📘 Lecture 29 — Concept of Total Derivatives
📖 Overview: This lecture introduces the concept of total differentiation, which captures the overall rate of change of a function when the independent variable affects the dependent variable through both direct and indirect channels. Understanding total derivatives is essential for analyzing complex economic relationships where variables are interdependent.
🗂️ Topics Covered
The lecture begins with the fundamental concept of total derivative, explaining the channel map and the decomposition into direct and indirect effects. It then applies total differentiation to analyze complementarity between coffee and sugar. Finally, it extends the concept to general and specific production functions with time-dependent labor and capital.
📝 Lecture Summary
Topic 138: Concept of Total Derivative
Assume y = f(x) where x = g(t). This creates a channel map: t causes x, and x causes y. When combining the two functions, we get y = f[g(t)]. Here, t is the ultimate source of change.
There are two channels through which t affects y:
- Indirect: via function
g. This meansx = g(t) ⇒ dx/dtand then(∂y/∂x)(dx/dt). - Direct: via function
f. This meansy = f(x)only iftappears directly inf, giving∂y/∂t.
The Total derivative (Direct + Indirect) of y with respect to t is:
dy/dt = ∂y/∂t + (∂y/∂x)(dx/dt)
🔑 Definition — Total Derivative: The total derivative dy/dt measures the total rate of change of y with respect to t, accounting for both the direct effect of t on y and the indirect effect through intermediate variable x.
📐 Formula: dy/dt = ∂y/∂t + (∂y/∂x)(dx/dt)
∂y/∂t= direct effect oftony(∂y/∂x)(dx/dt)= indirect effect viax
Caveat: If t does not appear directly in y = f(x), then ∂y/∂t = 0, and the total derivative simplifies to dy/dt = (∂y/∂x)(dx/dt).
📌 Example: Given y = 3x² + 5x + 2 and x = 2t² + 3, find dy/dt.
- Step 1: Find
dy/dx = 6x + 5 - Step 2: Find
dx/dt = 4t - Step 3: Apply total derivative formula:
dy/dt = (6x + 5)(4t) = 4t(6x + 5) = 24tx + 20t - Step 4: Substitute
x = 2t² + 3:dy/dt = 24t(2t² + 3) + 20t = 48t³ + 72t + 20t = 48t³ + 92t
Verification by substitution:
- Substitute
x = 2t² + 3intoy: y = 3(2t² + 3)² + 5(2t² + 3) + 2 y = 3(4t⁴ + 12t² + 9) + 10t² + 15 + 2y = 12t⁴ + 36t² + 27 + 10t² + 17y = 12t⁴ + 46t² + 44- Differentiating w.r.t.
t: dy/dt = 48t³ + 92t(Same as total derivative formula result) ✅
Topic 139: Complementarity Between Coffee and Sugar Using Total Derivative
Assume Q = f(C, S) where C is coffee and S is sugar. Furthermore, S = g(C). This implies complementarity between sugar and coffee.
Channel map: C causes S (via g function, specifically S = g(C)), and both C and S cause Q. The main variable causing change is C.
The total derivative formula for this relationship is:
dQ/dC = ∂Q/∂C + (∂Q/∂S)(dS/dC)
📌 Example: Given Q = 3C² + 2S² + 5CS and S = C² + 2C, find dQ/dC.
- Step 1: Find
∂Q/∂C = 6C + 5S - Step 2: Find
∂Q/∂S = 4S + 5C - Step 3: Find
dS/dC = 2C + 2 - Step 4: Apply formula:
dQ/dC = (6C + 5S) + (4S + 5C)(2C + 2) - Step 5: Substitute
S = C² + 2C:dQ/dC = [6C + 5(C² + 2C)] + [4(C² + 2C) + 5C](2C + 2)dQ/dC = (6C + 5C² + 10C) + (4C² + 8C + 5C)(2C + 2)dQ/dC = (5C² + 16C) + (4C² + 13C)(2C + 2)dQ/dC = 5C² + 16C + 8C³ + 8C² + 26C² + 26CdQ/dC = 8C³ + 39C² + 42C
Verification using substitution: Substitute S = C² + 2C into Q and differentiate directly with respect to C.
Q = 3C² + 2(C² + 2C)² + 5C(C² + 2C)Q = 3C² + 2(C⁴ + 4C³ + 4C²) + 5C³ + 10C²Q = 3C² + 2C⁴ + 8C³ + 8C² + 5C³ + 10C²Q = 2C⁴ + 13C³ + 21C²dQ/dC = 8C³ + 39C² + 42C✅ (Same result)
Topic 140: General Production Function with Time-Dependent Labor and Capital
This section extends the total derivative concept to a general production function. Consider a production function Q = f(L, K) where both L (labor) and K (capital) are functions of time t: L = h(t) and K = g(t).
The channel map shows that t affects both L and K, and both L and K affect Q. To find the total derivative of Q with respect to t:
dQ/dt = (∂Q/∂L)(dL/dt) + (∂Q/∂K)(dK/dt)
Since t does not appear directly in Q = f(L, K), there is no ∂Q/∂t term. The total derivative formula captures how output changes over time due to simultaneous changes in both labor and capital.
Topic 141: Specific Production Function with Time-Dependent Labor and Capital
This section applies the total derivative to a specific production function. Given Q = 3L² + 2K² + 5LK, where L = t² and K = 2t³, find dQ/dt.
- Step 1: Find
∂Q/∂L = 6L + 5K - Step 2: Find
∂Q/∂K = 4K + 5L - Step 3: Find
dL/dt = 2t - Step 4: Find
dK/dt = 6t² - Step 5: Apply formula:
dQ/dt = (6L + 5K)(2t) + (4K + 5L)(6t²) - Step 6: Substitute
L = t²andK = 2t³:dQ/dt = [6(t²) + 5(2t³)](2t) + [4(2t³) + 5(t²)](6t²)dQ/dt = (6t² + 10t³)(2t) + (8t³ + 5t²)(6t²)dQ/dt = 12t³ + 20t⁴ + 48t⁵ + 30t⁴dQ/dt = 12t³ + 50t⁴ + 48t⁵
Verification by substitution: Substitute L = t² and K = 2t³ into Q and differentiate with respect to t.
Q = 3(t²)² + 2(2t³)² + 5(t²)(2t³)Q = 3t⁴ + 2(4t⁶) + 10t⁵Q = 3t⁴ + 8t⁶ + 10t⁵dQ/dt = 12t³ + 48t⁵ + 50t⁴ = 12t³ + 50t⁴ + 48t⁵✅ (Same result)
💡 Why this matters: In economics, production functions rarely depend on a single variable. Total differentiation allows economists to model how output changes over time when both inputs (labor and capital) evolve simultaneously, which is essential for dynamic economic analysis.
⭐ Key Takeaways
The total derivative captures the complete rate of change of a function when an independent variable affects the dependent variable through multiple channels. The formula dy/dt = ∂y/∂t + (∂y/∂x)(dx/dt) decomposes change into direct and indirect components, with the caveat that the direct term vanishes if the independent variable does not appear explicitly. This concept extends naturally to multiple intermediate variables, as seen in production functions with time-dependent labor and capital. The total derivative can always be verified by direct substitution of the intermediate functions followed by ordinary differentiation, providing a powerful consistency check. Understanding total derivatives is fundamental for analyzing dynamic economic relationships such as input complementarity and production over time.
🧠 Quick Revision Questions
- What are the two channels through which a variable can affect a dependent function in total differentiation?
- State the total derivative formula for
y = f(x)wherex = g(t). When does the direct term disappear? - In the coffee-sugar complementarity example, what economic relationship does the total derivative
dQ/dCcapture? - Write the total derivative formula for a production function
Q = f(L, K)where bothLandKare functions of timet. - How can you verify the correctness of a total derivative calculation?
📘 Lecture 30 — Concept of Implicit Differentiation and Their Economic Applications
📖 Overview: This lecture introduces the concept of implicit differentiation, a technique for finding derivatives of functions that are not explicitly solved for one variable. It then applies this powerful method to derive key economic concepts, including marginal products, the marginal rate of technical substitution (MRTS), marginal utilities, and the marginal rate of substitution (MRS) from implicit production and utility functions. Finally, it demonstrates the technique on the Nerlove-Ringstad production function and a three-input logarithmic production function.
🗂️ Topics Covered
The lecture begins with the fundamental concept and method of implicit differentiation. It then applies this technique to analyze production functions, specifically deriving marginal physical products of labor (MPP_L) and capital (MPP_K), and the Marginal Rate of Technical Substitution (MRTS). The same approach is used to analyze utility functions, deriving marginal utilities (MU_1, MU_2, etc.) and the Marginal Rate of Substitution (MRS). The lecture concludes with applications to the Nerlove-Ringstad production function and a three-input logarithmic production function used in endogenous growth models.
📝 Lecture Summary
TOPIC 142: CONCEPT OF IMPLICIT DIFFERENTIATION
This topic deals with implicit functions, where the dependent variable is not expressed solely in terms of the independent variable. To find the derivative (dy/dx) of an implicit function, two steps are followed: 1) Differentiate each side of the equation with respect to (x), treating (y) as a function of (x), and 2) Solve the resulting equation for (dy/dx).
🔑 Definition — Implicit Function: A function where the relationship between the variables is given by an equation, not as a variable explicitly solved for another (e.g., (F(x, y) = 0)).
📐 Example: Find (dy/dx) for the implicit function (x^2 + y^2 = 4).
- Differentiate w.r.t. (x): (2x + 2y \frac{dy}{dx} = 0)
- Solve for (\frac{dy}{dx}): (2y \frac{dy}{dx} = -2x \implies \frac{dy}{dx} = -\frac{x}{y}) The result is the implicit derivative of the given implicit function.
TOPIC 143: PRODUCTION FUNCTION ANALYSIS USING IMPLICIT DIFFERENTIATION
Assume a production function (Q = f(K, L)). Its implicit function version is (F(Q, K, L) = f(K, L) - Q = 0). Marginal physical products of labor (MPP_L) and capital (MPP_K) are the partial derivatives (\frac{\partial Q}{\partial L}) and (\frac{\partial Q}{\partial K}), respectively. These can be found using implicit differentiation.
-
For MPP_K ((Q_K)), assume (Q = \bar{Q}) (constant).
- Writing implicit function: (F(K, L, \bar{Q}) = f(K, L) - \bar{Q} = 0).
- Using implicit differentiation: (\frac{\partial Q}{\partial K} = -\frac{F_K}{F_Q}).
- Result: (\frac{\partial Q}{\partial K} = \frac{f_K}{1} = f_K).
- The marginal physical product of capital ((MPP_K)) is expressed in relation to the function (f(K, L)).
-
For MPP_L ((Q_L)), assume (Q = \bar{Q}).
- Using implicit differentiation: (\frac{\partial Q}{\partial L} = -\frac{F_L}{F_Q}).
- Result: (\frac{\partial Q}{\partial L} = \frac{f_L}{1} = f_L).
- The marginal physical product of labor ((MPP_L)) is expressed in relation to the function (f(K, L)).
-
For the Marginal Rate of Technical Substitution ((MRTS_{LK})), found using implicit differentiation.
- Result: (MRTS_{LK} = \frac{dK}{dL} = -\frac{f_L}{f_K}).
- The marginal rate of technical substitution ((MRTS)) is expressed in relation to the function (f(K, L)).
TOPIC 144: MARGINAL RATE OF TECHNICAL SUBSTITUTION USING IMPLICIT DIFFERENTIATION
Assume (F(K, L) = 0) is an implicit function that can yield a production function: (Q = f(K, L)). The marginal rate of technical substitution, (MRTS_{LK} = \frac{dK}{dL}) along an isoquant. Using implicit differentiation:
- Assume (Q = \bar{Q}), implicit function: (f(K, L) - \bar{Q} = 0).
- (\frac{dK}{dL} = -\frac{f_L}{f_K})
- Result: The marginal rate of technical substitution ((MRTS_{LK})) is expressed in relation to the function (f(K, L)).
TOPIC 145: MARGINAL UTILITIES AND MARGINAL RATE OF SUBSTITUTION USING IMPLICIT DIFFERENTIATION
Assume an implicit equation: (U(x_1, x_2, ..., x_n) = c). A utility function can be extracted: (c = U(x_1, x_2, ..., x_n)).
🔑 Definition — Marginal Utility (MU): The change in total utility resulting from a one-unit change in the consumption of a good, holding all other goods constant.
-
For (MU_1) ((U_1)), assume (c = U(x_1, x_2, ..., x_n)) and (dx_2 = dx_3 = ... = dx_n = 0) (ceteris paribus). The implicit function is (U(x_1, \bar{x_2}, ..., \bar{x_n}) - c = 0).
- Result: (\frac{\partial U}{\partial x_1} = U_1) is the marginal utility due to an additional unit of (x_1) (the 1st good).
-
For (MU_2) ((U_2)), assume (dx_1 = dx_3 = ... = dx_n = 0).
- Result: (U_2) is the marginal utility of the second good.
-
For the Marginal Rate of Substitution ((MRS_{12})), assume (U(x_1, x_2) = c).
- Result: (\frac{dx_2}{dx_1} = -\frac{U_1}{U_2} = MRS_{12}).
-
For a higher-dimensional case, the Marginal Rate of Substitution between good 1 and good 2 is found by taking the total differential and setting it to zero:
- (U_1 dx_1 + U_2 dx_2 = 0 \implies \frac{dx_2}{dx_1} = -\frac{U_1}{U_2}).
TOPIC 146: NERLOVE-RINGSTAD PRODUCTION FUNCTION USING IMPLICIT DIFFERENTIATION
This production function, attributed to Nerlove (1963) and Ringstad (1967), is: [ Y = { \alpha L^\rho + \beta K^\rho }^\nu ] Where (Y, L, K > 0), and (\alpha, \beta, \nu > 0). This is often written as (|Y| = |Y^\nu|). The function can be rearranged as: [ Y^{1/\nu} = \alpha L^\rho + \beta K^\rho ] Which is not in an explicit function form. We use implicit differentiation.
-
Implicit differentiation w.r.t. (L):
- The derivative (\partial Y / \partial L) is derived as:
- [ \frac{\partial Y}{\partial L} = \nu \alpha \rho Y \left[ \frac{Y^{(1-\rho)/\nu}}{\alpha L^\rho + \beta K^\rho} \right] L^{\rho-1} ]
- In simpler notation, this can be expressed as: (\frac{\partial Y}{\partial L} = \frac{\nu \alpha \rho L^{\rho-1}}{(\alpha L^\rho + \beta K^\rho)^{1-1/\nu}}).
-
Implicit differentiation w.r.t. (K):
- The derivative (\partial Y / \partial K) is derived similarly as:
- [ \frac{\partial Y}{\partial K} = \frac{\nu \beta \rho K^{\rho-1}}{(\alpha L^\rho + \beta K^\rho)^{1-1/\nu}} ]
TOPIC 147: MARGINAL PRODUCTS OF THREE INPUT LOGARITHMIC PRODUCTION FUNCTION
Assume an endogenous growth model: (Y = AK^\alpha L^\beta H^\gamma), where (H) is human capital. Becker (1964) defined human capital as skills and adequate motivation to apply them. By taking logarithms, the production function is linearized: [ \ln Y = \ln A + \alpha \ln K + \beta \ln L + \gamma \ln H ]
For the marginal product of labor ((MP_L)), we partially differentiate with respect to the logarithm of (L).
- (\frac{\partial \ln Y}{\partial \ln L} = \beta), which is the labor elasticity of output.
- The marginal product of labor, (MP_L = \frac{\partial Y}{\partial L} = \beta \frac{Y}{L}).
For the marginal product of capital ((MP_K)):
- (\frac{\partial \ln Y}{\partial \ln K} = \alpha), which is the capital elasticity of output.
- The marginal product of capital, (MP_K = \frac{\partial Y}{\partial K} = \alpha \frac{Y}{K}).
For the marginal product of human capital ((MP_H)):
- (\frac{\partial \ln Y}{\partial \ln H} = \gamma), which is the human capital elasticity of output.
- The marginal product of human capital, (MP_H = \frac{\partial Y}{\partial H} = \gamma \frac{Y}{H}).
⭐ Key Takeaways
Implicit differentiation is a fundamental tool for deriving key economic concepts directly from implicit functions without needing to solve for one variable explicitly. In production analysis, it shows that marginal physical products ((MPP_L) and (MPP_K)) are simply the partial derivatives of the production function, while the Marginal Rate of Technical Substitution ((MRTS_{LK})) is the negative ratio of these partials. For utility functions, marginal utilities ((MU_i)) are partial derivatives, and the Marginal Rate of Substitution ((MRS_{ij})) is the negative ratio of the marginal utilities of the two goods. Finally, for a Cobb-Douglas-type production function in log-linear form, the coefficients ((\alpha, \beta, \gamma)) directly represent the elasticities of output with respect to each input, making implied marginal products easy to calculate (e.g., (MP_L = \beta Y/L)).
🧠 Quick Revision Questions
- What are the two steps to find (dy/dx) for an implicit function (F(x, y) = 0)?
- How is the Marginal Rate of Technical Substitution ((MRTS_{LK})) related to the marginal physical products of labor and capital?
- Using implicit differentiation, prove that the Marginal Rate of Substitution ((MRS_{12})) equals (-MU_1 / MU_2) for a utility function (U(x_1, x_2)).
- For the Nerlove-Ringstad production function, what is the expression for (\partial Y / \partial K) derived via implicit differentiation?
- In the three-input log-linear production function (\ln Y = \ln A + \alpha \ln K + \beta \ln L + \gamma \ln H), how would you calculate the marginal product of physical capital ((MP_K))?
📘 Lecture 31 — Exponential Functions and Growth
📖 Overview: This lecture introduces exponential functions and their application to growth processes in economics, particularly focusing on continuous compounding of interest. It covers the natural exponential function, instantaneous rates of growth, the relationship between continuous and discrete growth, discounting as negative growth, and practical applications of continuous compounding formulas.
🗂️ Topics Covered
The lecture covers exponential functions as representations of quantities that grow by fixed factors per time unit, the natural exponential function with base e, instantaneous rate of growth derived from continuous compounding value functions, numerical examples calculating instantaneous growth rates, comparisons between continuous and discrete growth patterns, discounting as negative growth transformations, and a practical application of continuous compounding to calculate investment balances over time.
📝 Lecture Summary
Topic 148: Exponential Functions and Growth
A quantity that increases (or decreases) by a fixed factor per unit of time is said to increase (or decrease) exponentially. The general form is:
V(t) = V₀bᵗ
Where b is the factor by which V increases when t increases by 1. Each base b gives a different value of V(t) = V₀bᵗ (e.g., b = 2, 3, etc.).
In calculus, most exponential functions appear with base e, which is an irrational number (approximately 2.71828). The function V(t) = V₀eᵗ is known as the natural exponential function. This function exemplifies growth of a sum of money capital over time and can also represent growth of population, wealth, or real capital.
Topic 149: Instantaneous Rate of Growth
Given a value function of continuous interest compounding:
V = Aeʳᵗ
Where: t = points in time, A = principal amount, e = natural base, r = interest rate, V = value at point in time.
The instantaneous rate of change is found by taking the derivative:
dV/dt = d(Aeʳᵗ)/dt = Aeʳᵗ * r = rV
For the instantaneous growth rate, we find the rate of change of value in relative (%) terms:
(dV/dt)/V = rV/V = r
The instantaneous rate of growth is time-dependent in general but is constant r for simplicity.
Numerical Example: Given V(t) = 100e⁰·⁰⁵ᵗ, the instantaneous rate of growth of V is 0.05 — a constant.
💡 Why this matters: The instantaneous rate of growth in continuous compounding equals the interest rate r itself, showing that the percentage growth rate is constant over time.
Topic 150: Numerical Examples of Instantaneous Rate of Growth
Example 1: Given V(t) = e⁰·⁰⁵ᵗ
Interest compounding: V = Aeʳᵗ where A = 1 (principal amount), r = 0.05 (instantaneous growth rate)
dV/dt = d(e⁰·⁰⁵ᵗ)/dt = e⁰·⁰⁵ᵗ * 0.05 = 0.05e⁰·⁰⁵ᵗ
(dV/dt)/V = (0.05e⁰·⁰⁵ᵗ)/(e⁰·⁰⁵ᵗ) = 0.05
Example 2: Given V(t) = 0.03e⁰·⁰⁶ᵗ
Interest compounding: V = Aeʳᵗ where A = 0.03 (principal amount), r = 0.06 (instantaneous growth rate)
dV/dt = d(0.03e⁰·⁰⁶ᵗ)/dt = 0.03e⁰·⁰⁶ᵗ * 0.06 = 0.0018e⁰·⁰⁶ᵗ
(dV/dt)/V = (0.0018e⁰·⁰⁶ᵗ)/(0.03e⁰·⁰⁶ᵗ) = 0.06
📌 Example: The instantaneous growth rate remains constant at r regardless of the principal amount A. For V(t) = 0.03e⁰·⁰⁶ᵗ, the growth rate is 0.06 or 6%.
Topic 151: Continuous vs Discrete Growth
Usually in economic situations, growth does not always take place on a continuous basis, not even in interest compounding. However, for discrete growth, where changes occur only once per period rather than from instant to instant, the assumption of continuous exponential growth function can be justified.
For discrete compounding: A(1+r)ᵗ where the series is A(1+r), A(1+r)², A(1+r)³, A(1+r)ⁿ
Here, exponents are time periods covered in compounding, A is principal amount, and r is interest rate. This series can be considered an exponential expression as Abᵗ, meaning V = Abᵗ even for discrete t if b = 1+r in % terms.
To convert base to natural exponent, compare Abᵗ = A(1+r)ᵗ with standard form of natural exponential function Aeʳᵗ:
A(1+r)ᵗ = A(e^(ln(1+r)))ᵗ = Ae^(t·ln(1+r))
This means ln(1+r) = r̃, where r̃ is the continuous growth rate, and for small r, ln(1+r) ≈ r.
This is why natural exponential functions are extensively applied in economic analysis despite not all growth patterns being purely continuous.
Topic 152: Discounting and Negative Growth
From interest compounding to discounting: The compound-interest problem finds future value V from present value A. The discounting problem finds present value A of a given sum V available after t years.
Discrete case: V = A(1+r)ᵗ
Cross-multiplying: A = V/(1+r)ᵗ = V(1+r)^(-t)
(Note the reversed roles of A and V compared to the introduction.)
Continuous case: Principal A grows into V = Aeʳᵗ after t years of continuous compounding at rate r.
Cross-multiplying: A = V/eʳᵗ = Ve^(-rt)
Here, e^(-rt) is also referred to as the discounting factor. This transforms growth into negative-growth.
🔑 Definition — Discounting factor (continuous): e^(-rt), where r is the interest rate and t is time. It represents the factor by which a future value must be multiplied to find its present value.
Topic 153: Applications of Continuous Compounding
Suppose the sum of PKR 5000 is invested in an account earning interest at an annual rate of 9%. What will be the balance after 8 years if interest is compounded continuously?
The formula for continuous compounding is: V = Aeʳᵗ
Here, A = PKR 5000, r = 0.09 (9%), t = 8 years.
Substituting values: V = 5000e^(0.09 × 8) V = 5000e⁰·⁷² V ≈ 5000 × 2.05443 V ≈ PKR 10,272.17
📌 Example: Principal will increase from PKR 5000 to PKR 10,272.17 after 8 years, if the annual interest is 9% and there is continuous compounding of interest.
⭐ Key Takeaways
The natural exponential function V(t) = Aeʳᵗ models continuous growth where the instantaneous rate of growth equals the constant r, found by dividing the derivative dV/dt by V itself. For discrete compounding A(1+r)ᵗ, the base can be converted to natural exponent using A(1+r)ᵗ = Ae^(t·ln(1+r)), allowing continuous analysis of discrete growth patterns. Discounting reverses the compounding process: while compounding finds future value from present value (V = Aeʳᵗ), discounting finds present value from future value (A = Ve^(-rt)), where e^(-rt) serves as the discounting factor representing negative growth. In the practical application, PKR 5000 compounded continuously at 9% for 8 years grows to approximately PKR 10,272.17.
🧠 Quick Revision Questions
- What is the formula for the natural exponential function representing continuous compounding, and what does each variable represent?
- How do you calculate the instantaneous rate of growth from a continuous compounding value function, and why is it constant?
- What is the relationship between discrete growth A(1+r)ᵗ and continuous growth Aeʳᵗ, and how can one be converted to the other?
- What is the discounting factor in continuous compounding, and how does discounting transform the role of present value A and future value V?
- If PKR 5000 is invested at 9% annual interest compounded continuously for 8 years, what is the final balance? Show your calculation.
📘 Lecture 32 — Use of Logarithms in Economics
📖 Overview: This lecture introduces logarithms as the inverse of exponentiation, explains their mathematical properties, and demonstrates their application in transforming production functions like the Cobb-Douglas and CES functions. Understanding logarithms is crucial for linearizing nonlinear economic relationships and simplifying estimation.
🗂️ Topics Covered
The lecture covers the meaning and types of logarithms (common and natural logs), the three fundamental laws of logarithms (product, quotient, and power rules) with economic examples, and the application of these laws to transform a CES production function, including a first- and second-order approximation of log(1 + z).
📝 Lecture Summary
TOPIC 154: LOGARITHMS MEANING AND TYPES
Attributed to a Scottish mathematician John Napier (1550-1617). The etymology is from Greek "logos" (ratio) + "arithmos" (number). A logarithm is the inverse of exponentiation. It helps to find the power of a number whose result is known.
If we have a^b = c, then log_a(c) = b.
Logarithms can have a base different than 10, such as the natural logarithm with base e ≈ 2.718 (Euler’s number). ln(x) is the natural logarithm of x. Contrary to the natural exponential function e^x, ln(x) and e^x are inverse functions of each other. This relationship demonstrates order independence in inverse, meaning ln(e^x) = x and e^(ln x) = x.
TOPIC 155: LAWS OF LOGARITHMS
1 - Logarithm of a Product
The log of a product equals the sum of the logs.
log_a(X · Y) = log_a(X) + log_a(Y)
Or equivalently: ln(X · Y) = ln(X) + ln(Y)
🔑 Definition — Log of a Product: log_a(XY) = log_a(X) + log_a(Y)
📐 Formula: ln(XY) = ln(X) + ln(Y) → The log of multiplied numbers is the sum of their individual logs.
📌 Example: log_2(8 × 4) = log_2(8) + log_2(4). We know log_2(8) = 3 and log_2(4) = 2. So log_2(32) = 3 + 2 = 5. Using a calculator: log_2(32) = 5 ✔️
💡 Why this matters: This is used to decompose multiplicative economic relationships. Economic Instance: ln(P · Q) = ln(P) + ln(Q) where P is price and Q is quantity, allowing us to separate the log of total revenue into its components.
2 - Logarithm of a Quotient
The log of a quotient equals the difference of the logs.
log_a(X/Y) = log_a(X) - log_a(Y)
OR ln(X/Y) = ln(X) - ln(Y)
🔑 Definition — Log of a Quotient: log_a(X/Y) = log_a(X) - log_a(Y)
📐 Formula: ln(X/Y) = ln(X) - ln(Y) → The log of a division is the subtraction of the logs.
📌 Example: log_2(8/4) = log_2(8) - log_2(4). We know log_2(8) = 3 and log_2(4) = 2. So log_2(2) = 3 - 2 = 1. Using a calculator: log_2(2) = 1 ✔️
💡 Why this matters: This is useful for analyzing ratios in economics. Economic Instance: ln(MP_K / MP_L) = ln(MP_K) - ln(MP_L) where MP is marginal product, used to analyze the ratio of factor productivities.
3 - Logarithm of a Power
The log of a number raised to a power is the power times the log of the number.
log_a(X^C) = C · log_a(X)
OR ln(X^C) = C · ln(X)
🔑 Definition — Log of a Power: log_a(X^C) = C · log_a(X)
📐 Formula: ln(X^C) = C · ln(X) → Exponents become coefficients in the logarithmic form.
📌 Example: log_2(8^3) = 3 · log_2(8). We know log_2(8) = 3. So log_2(512) = 3 · 3 = 9. Using a calculator: log_2(512) = 9 ✔️
💡 Why this matters: This is the key to linearizing power functions. Economic Instance: The log-linearized form of a Cobb-Douglas production function. Starting with Q = A · K^α · L^β, taking logs: ln(Q) = ln(A) + ln(K^α) + ln(L^β). Applying the product and power rules: ln(Q) = ln(A) + α · ln(K) + β · ln(L). This transforms a nonlinear multiplicative function into a linear equation in the logs of the variables, which is easy to estimate.
TOPIC 156: LAWS OF LOGARITHMS FOR TRANSFORMATION OF CES PRODUCTION FUNCTION
The Constant Elasticity of Substitution (CES) production function is given by:
Q = A [δ · K^(-ρ) + (1-δ) · L^(-ρ)]^(-ν/ρ)
Taking natural logs of both sides:
ln(Q) = ln(A) + (-ν/ρ) · ln[δ · K^(-ρ) + (1-δ) · L^(-ρ)]
Let Z = δ · K^(-ρ) + (1-δ) · L^(-ρ). So:
ln(Q) = ln(A) + (-ν/ρ) · ln(Z)
Since ρ is often small, we can take a first-order approximation of ln(1 + Z) using the series ln(1 + z) = z - z^2/2 + z^3/3 - .... The first-order approximation is ln(1+z) ≈ z. A second-order approximation is ln(1+z) ≈ z - z^2/2.
Setting Z as a small number and applying the approximation, we can linearize the CES function for estimation purposes, showing that it simplifies to forms similar to the Cobb-Douglas function under certain parameters.
⭐ Key Takeaways
Logarithms are the inverse of exponentiation, with base 10 (common) and base e (natural) being the most important types. The three laws of logs (product → sum, quotient → difference, power → multiplication) are essential tools for transforming multiplicative and exponential economic models into additive and linear forms. These laws are directly applied to linearize the Cobb-Douglas production function into ln(Q) = ln(A) + α·ln(K) + β·ln(L). For more complex functions like the CES, log laws combined with series approximations (like ln(1+z) ≈ z) provide a path to linearization and empirical estimation.
🧠 Quick Revision Questions
- What is the etymology of the word "logarithm" and who is credited with its invention?
- State and write the formula for the logarithm of a quotient. Provide an economic example of its use.
- Using the laws of logarithms, fully linearize the Cobb-Douglas function
Q = 5 · K^0.3 · L^0.7. - What is the first-order and second-order series approximation for
ln(1 + z)? - What key step involving the parameter
Zis taken to transform the CES production function using logarithms?
📘 Lecture 33 — RULES OF DIFFERENTIATION OF EXPONENTIAL AND LOGARITHMIC FUNCTIONS
📖 Overview: This lecture introduces the rules for differentiating exponential and logarithmic functions, which are essential for modeling growth processes in economics. It then applies these rules to solve several "optimal timing" problems, including wine storage, timber cutting, land speculation, art collection, and diamond purchase, demonstrating how to determine the optimal time to sell an asset for maximum present value.
🗂️ Topics Covered
The lecture begins with the differentiation rules for exponential functions [d/dx e^(f(x)) = e^(f(x)) * f'(x)] and logarithmic functions [d/dx ln(f(x)) = f'(x)/f(x)], illustrated by examples. It then applies these rules to a series of optimal timing problems, including a wine storage problem with value V(t)=Ke^(√t) and discount rate ρ. The lecture extends the framework to timber cutting, land purchase for speculation, art collection, and diamond purchase, each with its own value function and discount rate, solved using first-order conditions from maximizing present value.
📝 Lecture Summary
TOPIC 157: RULES OF DIFFERENTIATION OF EXPONENTIAL AND LOGARITHMIC FUNCTIONS
This topic covers the rules for differentiating exponential functions and logarithmic functions.
🔑 Definition — Exponential Function Rule: The derivative of e^(f(x)) is e^(f(x)) * f'(x).
📐 Formula: d/dx [ e^(f(x)) ] = e^(f(x)) * f'(x)
📌 Example:
Given f(x) = x^2 + 3, then f'(x) = 2x.
The derivative is:
d/dx [ e^(x^2 + 3) ] = e^(x^2 + 3) * (2x)
🔑 Definition — Logarithmic Function Rule: The derivative of ln(f(x)) is f'(x)/f(x) (where f(x) > 0).
📐 Formula: d/dx [ ln(f(x)) ] = f'(x) / f(x)
📌 Example:
Given f(x) = x^2 + 3, then f'(x) = 2x.
The derivative is:
d/dx [ ln(x^2 + 3) ] = (2x) / (x^2 + 3)
TOPIC 158: OPTIMAL TIMING: A PROBLEM OF WINE STORAGE
This problem determines the optimal time to sell a bottle of wine that increases in value over time. The value of the wine after t years is given by V(t) = Ke^(√t). You can sell it now for K dollars or store it and sell it later for the higher value V(t). The objective is to maximize the present value (W(t)) of selling at time t.
🔑 Definition — Present Value: The value today of a future payment, discounted by the interest rate (ρ).
📐 Formula: W(t) = V(t) * e^(-ρt) = Ke^(√t) * e^(-ρt) = Ke^(√t - ρt)
Assumptions: Storage costs are zero, and the wine is a sunk cost (already owned).
The First-Order Condition (F.o.c.) requires setting the derivative of W(t) with respect to time to zero: dW/dt = 0. This is found by differentiating ln(W).
💡 Why this matters: Taking the log of W simplifies the differentiation of the exponential function.
ln(W) = ln(K) + √t - ρt
d/dt [ln(W)] = 1/(2√t) - ρ
Setting this to zero: 1/(2√t) - ρ = 0 => 1/(2√t) = ρ => t* = 1/(4ρ^2)
The optimal time to sell (t*) is 1/(4ρ^2).
📌 Example: If the discount rate ρ = 0.1 (10%), then t* = 1 / (4 * 0.1^2) = 1 / (4 * 0.01) = 1 / 0.04 = 25 years.
The Second-Order Condition (S.O.C.) confirms this is a maximum (d²W/dt² < 0).
The Rate of Growth of V is d/dt [ln V] = 1/(2√t). The optimal time occurs when the rate of growth of V equals the discount rate ρ.
TOPIC 159: OPTIMAL TIMING: A PROBLEM OF TIMBER CUTTING
This problem determines the optimal time to harvest timber that has already been planted. The value of the timber (in $1000) is given by V(t) = e^(√t). The goal is to maximize the present value W(t) = V(t) * e^(-ρt).
The analysis is mathematically identical to the wine storage problem, with V(t) = e^(√t). The first-order condition yields the same result: t* = 1/(4ρ^2).
Key insight: A greater discount rate (ρ) means the future is valued less, so the timber should be cut earlier.
📌 Example: If the discount rate ρ = 0.05 (5%), then t* = 1 / (4 * 0.05^2) = 100 years.
If ρ = 0.1 (10%), then t* = 1 / (4 * 0.1^2) = 25 years.
This shows that a higher discount rate leads to a shorter optimal growing period.
TOPIC 160: OPTIMAL TIMING: LAND PURCHASE FOR SPECULATION
This problem determines the optimal time to sell land purchased for speculation. The land's value increases according to V(t) = 1000 e^(√t). The present value is W(t) = 1000 e^(√t - ρt). Using the same logarithmic differentiation method:
ln(W) = ln(1000) + √t - ρt
d/dt [ln(W)] = 1/(2√t) - ρ
F.o.c.: 1/(2√t) - ρ = 0 => 1/(2√t) = ρ => √t* = 1/(2ρ) => t* = 1/(4ρ^2)
Evaluating with ρ = 0.1 (10%):
t* = 1/(4 * 0.1^2) = 1/(4 * 0.01) = 1/0.04 = 25 years.
The second-order condition confirms that this is a maximum (d²W/dt² < 0).
TOPIC 161: OPTIMAL TIMING: ART COLLECTION
This problem determines the optimal time to sell an art collection whose value is estimated as V(t) = 200000 (1 + t)^(√t). The present value is W(t) = 200000 (1 + t)^(√t) e^(-ρt).
To find the optimal t, take logs:
ln(W) = ln(200000) + √t * ln(1 + t) - ρt
Set the derivative equal to zero (F.o.c.). The solution yields t* = 15.24 years (given ρ = 0.1).
📌 Example: Taking the derivative of √t * ln(1+t) yields the F.o.c.: (1/(2√t))*ln(1+t) + √t*(1/(1+t)) = ρ. Solving this for t gives the optimal time of 15.24 years. This is the optimal time to sell the art collection.
TOPIC 162: OPTIMAL TIMING: DIAMOND PURCHASE
This problem determines the optimal time to sell a diamond bought for investment. The estimated value is V(t) = 250000 (1 + t)^(√t). The present value is W(t) = 250000 (1 + t)^(√t) e^(-ρt).
The analysis is structurally identical to the art collection problem. Taking logs and setting the F.o.c. to zero yields the result.
📌 Example: For a discount rate ρ = 0.1 (10%), the solution to the F.o.c. gives t* = 2.52 years. This means the diamond should be sold after 2.52 years.
⭐ Key Takeaways
The most critical concepts from this lecture are the explicit rules for differentiating exponential functions (d/dx e^(f(x)) = e^(f(x)) * f'(x)) and logarithmic functions (d/dx ln(f(x)) = f'(x)/f(x)). These rules are directly applied to solve optimal timing problems by maximizing the present value W(t) = V(t) * e^(-ρt). By taking the natural log of the present value, the differentiation simplifies, and the first-order condition d(ln W)/dt = 0 is used to find the optimal time t*. The fundamental economic principle is that the asset should be sold when its rate of growth equals the discount rate. The examples show how different value functions (from simple exponential to more complex exponential of a log) lead to different optimal times, and a higher discount rate (ρ) always leads to a shorter optimal holding period.
🧠 Quick Revision Questions
- What is the derivative of
e^(f(x))? - What is the derivative of
ln(f(x))? - In the wine storage problem, what is the present value formula
W(t)that we are trying to maximize? - State the key economic condition (the decision rule) that characterizes the optimal time to sell an asset in all these problems.
- If the discount rate
ρincreases, does the optimal time to sell an asset (like timber or wine) increase or decrease?
📘 Lecture 34 — Finding the Rate of Growth Using Exponential and Logarithmic Functions
📖 Overview: This lecture teaches how to calculate instantaneous growth rates using logarithmic differentiation. It demonstrates applications to exports, consumption, employment, sales, and profit, showing how exponential and logarithmic functions are essential tools for analyzing dynamic economic variables.
🗂️ Topics Covered
The lecture covers the fundamental formula for instantaneous growth rate using logarithmic differentiation, followed by applied examples including growth of exports of a country, point elasticity of demand, rates of growth of population and per capita consumption, per capita employment, export earnings, sales, and profit. Each topic demonstrates how to derive and compute growth rates from given functions and parameters.
📝 Lecture Summary
TOPIC 163: FINDING THE RATE OF GROWTH USING EXPONENTIAL AND LOGARITHMIC FUNCTIONS
Assume variable V depends on time t. The instantaneous growth rate is given by the logarithmic derivative:
[ \frac{dV/dt}{V} = \frac{d}{dt}[\ln V(t)] = \frac{V'(t)}{V(t)} ]
This formula shows that the rate of growth equals the derivative of the natural log of the function with respect to time.
📐 Formula: ( \frac{d}{dt}[\ln V(t)] = \frac{V'(t)}{V(t)} ) → The instantaneous growth rate is the logarithmic derivative.
📌 Example: If ( V(t) = V_0 e^{rt} ), then taking natural log: ( \ln V = \ln V_0 + rt ). Differentiating w.r.t. ( t ): ( \frac{d}{dt}[\ln V] = r ). Therefore, ( \frac{V'}{V} = r ). (Rate of growth of V is r)
💡 Why this matters: This logarithmic differentiation technique transforms multiplicative relationships into additive ones, simplifying growth rate calculations.
TOPIC 164: GROWTH OF EXPORTS OF A COUNTRY
A country exports goods ( G(t) ) and services ( S(t) ), both dependent on time. Total exports: ( X(t) = G(t) + S(t) ).
The rate of growth of exports is derived using the formula: [ \frac{\dot{X}}{X} = \frac{d}{dt}[\ln X(t)] = \frac{d}{dt}[\ln(G(t) + S(t))] ]
Using the chain rule: [ \frac{\dot{X}}{X} = \frac{G}{X} \cdot \frac{\dot{G}}{G} + \frac{S}{X} \cdot \frac{\dot{S}}{S} ]
This gives equation (A): ( \frac{\dot{X}}{X} = \frac{G}{X}\hat{G} + \frac{S}{X}\hat{S} ), where ( \hat{G} = \frac{\dot{G}}{G} ) and ( \hat{S} = \frac{\dot{S}}{S} ).
📌 Example: Given ( \hat{G} = 10% ) and ( \hat{S} = 5% ), and assuming initial shares ( \frac{G}{X} = \frac{1}{3} ) and ( \frac{S}{X} = \frac{2}{3} ): [ \frac{\dot{X}}{X} = \frac{1}{3}(10%) + \frac{2}{3}(5%) = \frac{10}{3} + \frac{10}{3} = \frac{20}{3}% \approx 6.67% ]
TOPIC 165: FINDING THE POINT ELASTICITY
Using the usual formula of price elasticity of demand: [ \epsilon = \frac{dQ}{dP} \cdot \frac{P}{Q} ]
In logarithmic form: ( \epsilon = \frac{d(\ln Q)}{d(\ln P)} )
📌 Example: Given demand function: ( Q = \frac{k}{P} ) or ( QP = k ). Taking logs: ( \ln Q + \ln P = \ln k ) Differentiating w.r.t. ( P ): [ \frac{d(\ln Q)}{d(\ln P)} + 1 = 0 ] [ \epsilon = \frac{d(\ln Q)}{d(\ln P)} = -1 ] Thus, ( \epsilon = -1 ) (unitary elastic demand curve).
TOPIC 166: RATES OF GROWTH OF POPULATION, CONSUMPTION, AND PER CAPITA CONSUMPTION
Growth rate of consumption ( C ) is ( \alpha ). Growth rate of population ( N ) is ( \beta ). Consumption per capita: ( c = \frac{C}{N} ).
Taking logs: ( \ln c = \ln C - \ln N ) Differentiating w.r.t. time: [ \frac{\dot{c}}{c} = \frac{\dot{C}}{C} - \frac{\dot{N}}{N} ]
Given ( \frac{\dot{C}}{C} = \alpha ) and ( \frac{\dot{N}}{N} = \beta ): [ \frac{\dot{c}}{c} = \alpha - \beta ]
🔑 Definition — Rate of growth of a quotient variable: The rate of growth of a quotient variable is equal to the difference of individual growth rates.
TOPIC 167: RATE OF GROWTH OF PER CAPITA EMPLOYMENT
Growth rate of employment opportunities ( E ) is ( \gamma ). Growth rate of population ( N ) is ( \delta ). Employment per capita: ( e = \frac{E}{N} ).
Using the same logarithmic differentiation: [ \frac{\dot{e}}{e} = \frac{\dot{E}}{E} - \frac{\dot{N}}{N} = \gamma - \delta ]
📌 Example: If ( \gamma = 5% ) and ( \delta = 2% ): [ \frac{\dot{e}}{e} = 5% - 2% = 3% ]
Growth rate of per capita employment equals difference of growth rate of employment and population.
TOPIC 168: RATE OF GROWTH OF EXPORT EARNINGS OF A COUNTRY
Two exports of a country: C and B. ( C = C_0 e^{0.10t} ) and ( B = B_0 e^{0.20t} ). C grows at 10% and B grows at 20%.
Total export earnings: ( X = C + B ) [ \frac{\dot{X}}{X} = \frac{C}{X} \cdot \frac{\dot{C}}{C} + \frac{B}{X} \cdot \frac{\dot{B}}{B} ]
Given ( \frac{\dot{C}}{C} = 0.10 ) and ( \frac{\dot{B}}{B} = 0.20 ): [ \frac{\dot{X}}{X} = \frac{C}{X}(0.10) + \frac{B}{X}(0.20) ]
📌 Example: If initial ( C = 40% ) and ( B = 60% ) of X: [ \frac{\dot{X}}{X} = 0.40(0.10) + 0.60(0.20) = 0.04 + 0.12 = 0.16 = 16% ]
TOPIC 169: RATE OF GROWTH OF SALES
Sales function: ( S(t) = 100 + \frac{240}{\sqrt{t}} ) [ S(t) = 100 + 240t^{-1/2} ]
Taking derivative: [ S'(t) = 0 + 240 \left(-\frac{1}{2}\right) t^{-3/2} = -120t^{-3/2} ]
Rate of growth of sales: [ \frac{S'}{S} = \frac{-120t^{-3/2}}{100 + 240t^{-1/2}} = \frac{-120}{t^{3/2}(100 + 240t^{-1/2})} ]
Simplifying: ( \frac{S'}{S} = \frac{-120}{100t^{3/2} + 240t} )
📌 Example: Assuming ( t = 4 ): [ \frac{S'}{S} = \frac{-120}{100(4)^{3/2} + 240(4)} = \frac{-120}{100(8) + 960} = \frac{-120}{800 + 960} = \frac{-120}{1760} = -0.0682 = -6.82% ]
💡 Why this matters: This shows sales are declining at 6.82% when ( t = 4 ). The negative growth rate indicates the sales function approaches its asymptote of 100.
TOPIC 170: RATE OF GROWTH OF PROFIT
In addition to maximization of profit, the rate of growth of profit can also be of interest for a firm. Profit function: ( \pi(t) = 100 + \frac{800}{\sqrt{t}} ) [ \pi(t) = 100 + 800t^{-1/2} ]
Taking derivative: [ \pi'(t) = 800\left(-\frac{1}{2}\right) t^{-3/2} = -400t^{-3/2} ]
Rate of growth of profit: [ \frac{\pi'}{\pi} = \frac{-400t^{-3/2}}{100 + 800t^{-1/2}} = \frac{-400}{100t^{3/2} + 800t} ]
📌 Example: Assuming ( t = 8 ): [ \frac{\pi'}{\pi} = \frac{-400}{100(8)^{3/2} + 800(8)} = \frac{-400}{100(22.627) + 6400} = \frac{-400}{2262.7 + 6400} = \frac{-400}{8662.7} = -0.0462 = -4.62% ]
💡 Why this matters: When ( t = 8 ), profit is declining at approximately 4.62%, showing that even with positive profit levels, the growth rate may be negative.
⭐ Key Takeaways
The instantaneous growth rate of any variable is found by taking the natural logarithm and differentiating with respect to time, which transforms multiplicative relationships into additive ones. For a sum of variables, the overall growth rate is a weighted average of individual growth rates, with weights being the relative shares. For a quotient variable like per capita consumption, the growth rate equals the growth rate of the numerator minus the growth rate of the denominator. Logarithmic differentiation provides a direct method for calculating price elasticity of demand as the derivative of log quantity with respect to log price. These growth rate formulas apply to any economic variable, including exports, sales, and profit, allowing analysts to understand dynamic behavior beyond simple maximization.
🧠 Quick Revision Questions
-
What is the formula for instantaneous growth rate using logarithmic differentiation?
-
If total exports ( X = G + S ), and G grows at 8% while S grows at 12%, with shares ( \frac{G}{X} = 0.3 ) and ( \frac{S}{X} = 0.7 ), what is the growth rate of total exports?
-
How is the price elasticity of demand expressed in logarithmic form?
-
If consumption grows at 6% and population grows at 2%, what is the growth rate of per capita consumption?
-
Given sales function ( S(t) = 100 + 240t^{-1/2} ), calculate the growth rate when ( t = 9 ).