Econometrics Formulas

Paired t Test

𝑑̅√𝑛 𝛴𝑑 𝛴𝑑 2 − 𝑛(𝑑̅ )2
𝒕= ̅=
𝒅 𝒔𝒅 = √
𝑆𝑑 𝑛 𝑛−1

ANOVA

𝛴𝑥
𝐺𝑀 (𝑋ധ) =
𝑁
𝐵𝑒𝑡𝑤𝑒𝑒𝑛 𝑉𝑎𝑟𝑖𝑎𝑛𝑐𝑒
𝐹= 𝑊𝑖𝑡ℎ𝑖𝑛 𝑉𝑎𝑟𝑖𝑎𝑛𝑐𝑒

∑(𝑌1 − 𝑌ത1 )² ∑(𝑌2 − 𝑌ത2 )² ∑(𝑌3 − 𝑌ത3 )²

𝑇𝑜𝑡𝑎𝑙 𝑉𝑎𝑟𝑖𝑎𝑛𝑐𝑒 = + +
𝑛−1 𝑛−1 𝑛−1
2 2 2
𝑛1 ൫𝑌ത1 − 𝑋ധ൯ + 𝑛2 ൫𝑌ത2 − 𝑋ധ൯ + 𝑛3 ൫𝑌ത3 − 𝑋ധ ൯
𝐵𝑒𝑡𝑤𝑒𝑒𝑛 𝑉𝑎𝑟𝑖𝑎𝑛𝑐𝑒 =
𝑘−1
2 2
(𝑠1 2 +𝑠2 2 +𝑠3 2 ) ∑൫𝑌1 −𝑌ത1 ൯ + ∑൫𝑌2 −𝑌ത2 ൯ + ∑(𝑌3 −𝑌ത3 )²
𝑊𝑖𝑡ℎ𝑖𝑛 𝑉𝑎𝑟𝑖𝑎𝑛𝑐𝑒 = or 𝑊𝑖𝑡ℎ𝑖𝑛 𝑉𝑎𝑟𝑖𝑎𝑛𝑐𝑒 = 𝑛−𝑘
𝑛−𝑘

R2adj= 1 – (MSE / MST)

Linear Regression

Y = β0 + β1x1+ β2x2 + β3x3 + β4x4 + β5x5 + uij (ε)

ෝ = 𝛽መ0 + 𝛽መ1 𝑥1 + 𝛽መ2 𝑥2 + 𝛽መ3 𝑥3 + 𝛽መ4 𝑥4 + 𝛽መ5 𝑥5 + 𝑈𝑖𝑗

𝒚

Where,

β0 = Constant

β1, β2, β3, β4, β5 = Coefficients

β1x1+ β2x2 + β3x3 = Deterministic or Systematic Component

Uij = Error or Residual or Stochastic

Simple/Two Variable LR:

෡ 𝟏 = 𝛴𝑥𝑦
𝜷 ෡ 𝟎 = 𝑌ത − 𝛽መ1 𝑋ത
𝜷 ෡ = 𝛽መ0 + 𝛽መ1 𝑥𝑖
𝒀
𝛴𝑥 2
 𝛴𝑒 2 ∑ ቀ𝛽෠ 1 × 𝑥𝑦ቁ
𝟐
𝒓 = 1− 2 = 𝒓 = ඥ𝑟 2 𝑬𝒓𝒓𝒐𝒓(𝒆𝟐 ) = ∑𝑦2 ൫1 − 𝑅2 ൯;
𝛴𝑦 𝛴𝑦 2

∑𝑒 2
𝑽𝒂𝒓𝒊𝒂𝒏𝒄𝒆 𝒐𝒇 𝒆𝟐 =
𝑁−2

෡ 𝟏 = (𝜎 ∗ 𝑒 2 ) 1 ෡ 𝟎 = (𝜎 ∗ 𝑒 2 ) ∑𝑋 2
𝑽𝒂𝒓𝒊𝒂𝒏𝒄𝒆 𝒐𝒇 𝜷 𝑽𝒂𝒓𝒊𝒂𝒏𝒄𝒆 𝒐𝒇 𝜷
൫∑𝑥 2 ൯ 𝑛 ൫∑𝑥 2 ൯

Multiple LR:

෡ = 𝛽መ0 + 𝛽መ1 𝑥1 + 𝛽መ2 𝑥2

𝒀

2
෡ 𝟏 = ൫∑𝑥1 𝑦൯ × ൫∑𝑥2 ൯ − ൫∑𝑥2 𝑦൯ × ൫∑𝑥1 𝑥2 ൯
𝜷
൫∑𝑥1 2 ൯ × ൫∑𝑥2 2 ൯ − (∑𝑥1 𝑥2 )2

෡ ൫∑𝑥2 𝑦൯ × ൫∑𝑥1 2 ൯ − ൫∑𝑥1 𝑦൯ × ൫∑𝑥1 𝑥2 ൯

𝜷𝟐 =
൫∑𝑥1 2 ൯ × ൫∑𝑥2 2 ൯ − (∑𝑥1 𝑥2 )2

෡ 𝟎 = 𝑌ത − 𝛽መ1 𝑋ത1 − 𝛽መ2 𝑋ത2

𝜷

൫𝛽መ1 × ∑𝑦𝑥1 ൯ + (𝛽መ2 × 𝑦𝑥2 )

𝑹𝟐 =
∑𝑦 2

(𝑛 − 1)
𝐴𝑑𝑗𝑢𝑠𝑡𝑒𝑑2 = 1 − ൫1 − 𝑅 2 ൯
𝑛−𝑘

∑𝑒 2
𝑬𝒓𝒓𝒐𝒓൫𝒆𝟐 ൯ = ∑𝑦 2 ൫1 − 𝑅 2 ൯; 𝑽𝒂𝒓𝒊𝒂𝒏𝒄𝒆 𝒐𝒇 𝒆𝟐 =
𝑁−𝐾

෡𝟏 = (𝜎 ∗ 𝑒 2 ) × (∑𝑥2 2 )
𝑽𝒂𝒓𝒊𝒂𝒏𝒄𝒆 𝒐𝒇 𝜷
൫∑𝑥1 2 ൯ × ൫∑𝑥2 2 ൯ − (∑𝑥1 𝑥2 )2

(𝜎 ∗ 𝑒 2 ) × (∑𝑥1 2 )
෡𝟐 =
𝑽𝒂𝒓𝒊𝒂𝒏𝒄𝒆 𝒐𝒇 𝜷
൫∑𝑥1 2 ൯ × ൫∑𝑥2 2 ൯ − (∑𝑥1 𝑥2 )2

S.E.B1 =

Chi-Square (𝝌𝟐 )

𝟐
∑(𝑂 − 𝐸)2
𝝌 = 𝑊ℎ𝑒𝑟𝑒, 𝑶 = 𝑂𝑏𝑠𝑒𝑟𝑣𝑒𝑑 𝑉𝑎𝑙𝑢𝑒; 𝑬 = 𝐸𝑥𝑝𝑒𝑐𝑡𝑒𝑑 𝑉𝑎𝑙𝑢𝑒
𝐸
 𝑅1 × 𝐶1 𝑅1 × 𝐶2 𝑅2 × 𝐶1 𝑅2 × 𝐶2
𝑬𝟏𝟏 = 𝑬𝟏𝟐 = 𝑬𝟐𝟏 = 𝑬𝟐𝟐 =
𝑇𝑜𝑡𝑎𝑙(𝑛) 𝑛 𝑛 𝑛

MLR with Matrix

ෝ =∝ +𝛽መ1 𝑥1 + 𝛽መ2 𝑥2 + 𝛽መ3 𝑥3 + 𝑈𝑖𝑗

𝒚

2
𝛽መ1 𝑥11 𝑥12 𝑥13 ∑𝑥1 ∑𝑥1 𝑥2 ∑𝑥1 𝑥3
መ
൦𝛽1 ൪ 𝑋 = ൥𝑥21 𝑥22 𝑥23 ൩ ′
𝑿 𝑿 = ൦∑𝑥1 𝑥2 ∑𝑥2 2 ∑𝑥2 𝑥3 ൪
𝑥31 𝑥32 𝑥33 2
𝛽መ1 ∑𝑥3 𝑥1 ∑𝑥2 𝑥3 ∑𝑥3

1 1
𝑨−𝟏 = 𝐴𝑑𝑗 𝐴 𝑂𝑟 (𝑿′ 𝑿)−𝟏 = 𝐴𝑑𝑗 (𝑋 ′ 𝑋)
ȁ𝐴ȁ ȁ𝑋 ′ 𝑋ȁ

1
෡ = (𝑋 ′ 𝑋)−1 × (𝑋 ′ 𝑌)
𝜷 = 𝐴𝑑𝑗 (𝑋 ′ 𝑋) × (𝑋 ′ 𝑌)
ȁ𝑋 ′ 𝑋ȁ

ෝ = 𝑌ത − 𝛽መ1 𝑋ത1 − 𝛽መ2 𝑋ത2 − 𝛽መ3 𝑋ത3

∝

𝛽መ1 ∑𝑥1 𝑦 + 𝛽መ2 ∑𝑥2 𝑦 + 𝛽መ3 ∑𝑥3 𝑦

𝑹𝟐 =
∑𝑦 2

MLR with Equation

∑ y1 = nBo + B1 ∑ x1 + B2 ∑ x2

∑ y x1 = Bo ∑ x1 + B1 ∑ x1^2 + B2 ∑ x1x2

∑ y x2 = Bo ∑ x2 + B1 ∑ x1x2 + B2 ∑ x2^2

Cobb Douglas:

𝐴𝐿∝ 𝐾𝛽 … … … + 𝜇

→ 𝒍𝒐𝒈𝒆 𝑨 = 𝛼𝑙𝑜𝑔𝑒 𝐿 + 𝛽𝑙𝑜𝑔𝑒 𝐾

෡ and 𝛽 = 𝜷
Here, 𝛼 = 𝜷 ෡ [ According to Sec B]
𝟐 𝟑

1. 𝑙𝑜𝑔𝑒 𝐴 = 𝑌ത − 𝛼𝑋ത2 − 𝛽𝑋ത3

Σ(𝑦𝑥2 ) Σ(𝑥3 2 )− Σ(𝑥2 𝑥3 ) Σ(𝑦𝑋3 )
2. ∝
ෝ=
Σ(𝑥2 2 ) Σ(𝑥3 2 )− Σ(𝑥2 𝑥3 )
Σ(𝑦𝑥3 ) Σ(𝑥2 2 ) − Σ(𝑥2 𝑥3 ) Σ(𝑦𝑋2 )
3. 𝛽መ =
Σ(𝑥2 2 ) Σ(𝑥3 2 ) − Σ(𝑥2 𝑥3 )

൫෌ 𝑋1 ൯(෌ 𝑋1 ) ൫෌ 𝑋1 ൯(∑ 𝑌) ൫෌ 𝑋2 ൯(∑ 𝑌)

4. ∑ 𝑥1 𝑥2 = ∑ 𝑋1 𝑋2 − ∑ 𝑥1 𝑦 = ∑ 𝑋1 𝑌 − ∑ 𝑥2 𝑦 = ∑ 𝑋1 𝑌 −
𝑛 𝑛 𝑛
 ( ∑ 𝑋1 2 )2 ( ∑ 𝑋2 2 )2
5. ∑ 𝑥1 2 = ∑ 𝑋1 2 − ∑ 𝑥2 2 = ∑ 𝑋2 2 −
𝑛 𝑛

𝑅𝑗2 /(𝑘−1)
6. F test =
(1−𝑅𝑗2 )/(𝑛−𝑘)

Multicollinearity
1 1
Tolerance𝑖 = VIF = 1 − 𝑅𝑗2 VIF𝑖 =
𝑖 1−𝑅𝑗2

𝑅𝑗2 /(𝑘 − 2)
F test =
(1 − 𝑅𝑗2 )/(𝑛 − 𝑘 + 1)

Heteroscedasticity
1 𝑛−𝑑
Omit Central Observation = 𝑑 = 5 × 𝑛 𝑠𝑢𝑏𝑠𝑎𝑚𝑝𝑙𝑒 𝑛𝑢𝑚𝑏𝑒𝑟 = 2

𝑛−𝑑−2𝑘 𝐸𝑆𝑆
𝑑𝑓 = 𝐹 = 𝐸𝑆𝑆2
2 1

DW Test:

Σ(𝑒𝑡 − 𝑒𝑡−1 )2
𝑑𝑘 =
Σ𝑒𝑡 2