Scribd
Upload
10 views

Week+1+-318

The document discusses the impact of corruption and terrorism on firm sales and trade, highlighting that a 10% increase in corruption can lead to varying effects on sales depending on firm size. It introduces econometrics, emphasizing the importance of quantifying relationships between variables using statistics and mathematics, and outlines key concepts such as variance, covariance, and correlation. The course will focus on using Stata for analysis, with applications in various economic research institutions.

Uploaded by

Mostafa Allam
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
10 views

Week+1+-318

The document discusses the impact of corruption and terrorism on firm sales and trade, highlighting that a 10% increase in corruption can lead to varying effects on sales depending on firm size. It introduces econometrics, emphasizing the importance of quantifying relationships between variables using statistics and mathematics, and outlines key concepts such as variance, covariance, and correlation. The course will focus on using Stata for analysis, with applications in various economic research institutions.

Uploaded by

Mostafa Allam
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 6

Week 1

Introduction
1. Important Findings in Economic Research
𝐶𝑜𝑟𝑟𝑢𝑝𝑡𝑖𝑜𝑛 → 𝐹𝑖𝑟𝑚 𝑆𝑎𝑙𝑒𝑠
10% ↑ 𝐶𝑜𝑟𝑟𝑢𝑝𝑡𝑖𝑜𝑛 → ↓ 𝐹𝑖𝑟𝑚 𝑆𝑎𝑙𝑒𝑠 2.2% (𝑆𝑀𝐸)
10% ↑ 𝐶𝑜𝑟𝑟𝑢𝑝𝑡𝑖𝑜𝑛 → ↑ 𝐹𝑖𝑟𝑚 𝑆𝑎𝑙𝑒𝑠 5.5% (𝐿𝑎𝑟𝑔𝑒 𝑎𝑛𝑑 𝑀𝑁𝐸)

𝑇𝑒𝑟𝑟𝑜𝑟𝑖𝑠𝑚 → 𝑇𝑟𝑎𝑑𝑒
1 𝑖𝑛𝑐𝑖𝑑𝑒𝑛𝑡 𝑇𝑒𝑟𝑟𝑜𝑟𝑖𝑠𝑚 → ↓ 𝑇𝑟𝑎𝑑𝑒 2.2%
2. What Econometrics?

𝑄𝑢𝑎𝑛𝑡𝑖𝑓𝑦
⏞
𝐸𝑐𝑜𝑛𝑜 + ⏞ 𝑀𝑒𝑡𝑟𝑖𝑐𝑠
The Goal is 𝑡𝑜 Measure (Quantify/Assess) the relationship between different
variables.
Econometrics relies on Statistics and Mathematics to quantify (test) a given
economic theory.

Econometrics concerns about Regression:


𝑋 → 𝑌 (𝑏𝑦 𝑎 𝑔𝑖𝑣𝑒𝑛 𝛽̂ )
𝐶𝑜𝑟𝑟𝑢𝑝𝑡𝑖𝑜𝑛 → 𝑆𝑎𝑙𝑒𝑠 (𝑏𝑦 𝑎 𝑔𝑖𝑣𝑒𝑛 2.2%)
𝑇𝑒𝑟𝑟𝑜𝑟𝑖𝑠𝑚 → 𝑇𝑟𝑎𝑑𝑒 (𝑏𝑦 𝑎 𝑔𝑖𝑣𝑒𝑛 2.2%)
𝐼𝑛𝑑𝑒𝑝𝑒𝑛𝑑𝑒𝑛𝑡 → 𝑅𝑒𝑠𝑝𝑜𝑛𝑠𝑒 (𝐷𝑒𝑝𝑒𝑛𝑑𝑒𝑛𝑡) (𝑏𝑦 𝑎 𝑔𝑖𝑣𝑒𝑛 𝑒𝑓𝑓𝑒𝑐𝑡)

3. The course will focus on Stata


𝐸𝑥𝑐𝑒𝑙 → 𝑀𝑎𝑡𝑙𝑎𝑏 → 𝑆𝑃𝑆𝑆 → 𝐸𝑣𝑖𝑒𝑤𝑠 → 𝑆𝑡𝑎𝑡𝑎, 𝑃𝑦𝑡ℎ𝑜𝑛, 𝑅

4. Places that ask for STATA:


• Afreximbank
• Economic Research Forum (ERF)
• World Bank
• European Bank for Reconstruction and Development
• Others.

Part I: Introduction and Statistical Background


• Variance.

∑(𝑋 − 𝑋̅)2
𝑉(𝑋) =
𝑁
Data 1
X (Profits) (𝑋 − 𝑋̅)2
1 2.25
2 0.25
3 0.25
4 2.25
𝑋̅ = 2.5 ∑(𝑋 − 𝑋̅)2 = 5

∑(𝑋 − 𝑋̅)2 5
𝑉(𝑋) = = = 1.25
𝑁 4
Data 2
X (Profits) (𝑋 − 𝑋̅)2
1 9
2 4
3 1
10 36
̅
𝑋=4 ∑(𝑋 − 𝑋̅)2 = 50

∑(𝑋 − 𝑋̅)2 50
𝑉(𝑋) = = = 12.5
𝑁 4

Conclusion: Firms in Data 1 have profits that are more similar compared to
Firms in Data 2. Having a Rule/theory for Data 1 is easy and expected
compared to Data 2.

For Example,
Data 1: When Tax increases, we expect that Profits of Firms declines by a
given amount.
Data 2: When Tax increases, it is hard to reach a general rule or a given
(specific) expected decline in profits.

𝐸𝑐𝑜𝑛𝑜𝑚𝑒𝑡𝑟𝑖𝑐𝑠 𝑎𝑙𝑤𝑎𝑦𝑠 𝑑𝑖𝑠𝑐𝑢𝑠𝑠 𝑎𝑏𝑜𝑢𝑡 𝑀𝑖𝑛𝑖𝑚𝑖𝑧𝑖𝑛𝑔 𝑉𝑎𝑟𝑖𝑎𝑛𝑐𝑒.


You should not remove an outlier (like 10 for example in Data 2) unless
you have a valid reason to report.

Last Note: We should divide by 𝑁 − 1 if we are in a context of sample:


∑(𝑋 − 𝑋̅)2
𝑉(𝑋) =
𝑁−1

• Covariance.

+
𝐶𝑜𝑣(𝑋, 𝑌) = { 0
−

∑(𝑋 − 𝑋̅)(𝑌 − 𝑌̅)


𝐶𝑜𝑣(𝑋, 𝑌) =
𝑁

𝑋 𝑌 𝑋 − 𝑋̅ 𝑌 − 𝑌̅ (𝑋 − 𝑋̅)(𝑌 − 𝑌̅)
1 2 -1.5 -3 4.5
2 4 -0.5 -1 0.5
3 6 0.5 1 0.5
4 8 1.5 3 4.5
Averages 2.5 5 Sum 10
∑(𝑋 − 𝑋̅)(𝑌 − 𝑌̅) 10
𝐶𝑜𝑣(𝑋, 𝑌) = = = 2.5
𝑁 4

𝑌𝑜𝑢 𝐶𝐴𝑁𝑁𝑂𝑇 interpret the magnitude (number) but you can interpret the
sign (+). We conclude that we have a positive relationship between the two
variables. We cannot conclude if this relationship is strong or weak.

• Correlation.

𝑋 𝑌 𝑋 − 𝑋̅ 𝑌 − 𝑌̅ (𝑋 − 𝑋̅)(𝑌 − 𝑌̅) (𝑋 − 𝑋̅)2 (𝑌 − 𝑌̅)2


1 2 -1.5 -3 4.5 2.25 9
2 4 -0.5 -1 0.5 0.25 1
3 6 0.5 1 0.5 0.25 1
4 8 1.5 3 4.5 2.25 9
Averages 2.5 5 Sum 10 5 20

+1
𝐶𝑜𝑟𝑟(𝑋, 𝑌) = { 0
−1

∑(𝑋 − 𝑋̅)(𝑌 − 𝑌̅)


𝐶𝑜𝑟𝑟(𝑋, 𝑌) =
𝑁. 𝜎𝑋 . 𝜎𝑌

𝜎𝑋 : 𝑆𝑡𝑎𝑛𝑑𝑎𝑟𝑑 𝐷𝑒𝑣𝑖𝑎𝑡𝑖𝑜𝑛 𝑜𝑓 𝑋 → √𝑉(𝑋)

∑(𝑋 − 𝑋̅)2 5
𝜎𝑋 = √ =√
𝑁 4

𝜎𝑌 : 𝑆𝑡𝑎𝑛𝑑𝑎𝑟𝑑 𝐷𝑒𝑣𝑖𝑎𝑡𝑖𝑜𝑛 𝑜𝑓 𝑌 → √𝑉(𝑌)


∑(𝑌 − 𝑌̅ )2 20
𝜎𝑋 = √ =√
𝑁 4

∑(𝑋 − 𝑋̅ )(𝑌 − 𝑌̅) 10


𝐶𝑜𝑟𝑟(𝑋, 𝑌) = = = +1
𝑁. 𝜎𝑋 . 𝜎𝑌 5 20
4. √ . √
4 4
We can interpret the sign and magnitude. The relationship between the two
variables is positive and Strong (Maximum).

𝑆𝑢𝑚 − 𝑢𝑝
(1) Variance

𝑉(𝑋) → 𝑆ℎ𝑜𝑢𝑙𝑑 𝑏𝑒 𝑎𝑡 𝑀𝑖𝑛𝑖𝑚𝑢𝑚 𝑡𝑜 𝑔𝑢𝑎𝑟𝑎𝑛𝑡𝑒𝑒 𝑡ℎ𝑎𝑡 𝑖𝑛𝑑𝑖𝑣𝑖𝑑𝑢𝑎𝑙𝑠


𝑖𝑛 𝑚𝑦 𝑑𝑎𝑡𝑎 ℎ𝑎𝑣𝑒 𝑡ℎ𝑒 𝑠𝑎𝑚𝑒 𝑏𝑒ℎ𝑎𝑣𝑖𝑜𝑟

(2) Covariance

→ 𝐼𝑚𝑝𝑜𝑟𝑡𝑎𝑛𝑡 𝑡𝑜 𝑘𝑛𝑜𝑤 𝑡ℎ𝑒 𝑟𝑒𝑙𝑎𝑡𝑖𝑜𝑛𝑠ℎ𝑖𝑝 𝑏𝑒𝑡𝑤𝑒𝑒𝑛𝑡ℎ𝑒 𝑡𝑤𝑜 𝑣𝑎𝑟𝑖𝑎𝑏𝑙𝑒𝑠


𝑤𝑒 𝑠ℎ𝑜𝑢𝑙𝑑 𝑛𝑜𝑡 𝑖𝑛𝑡𝑒𝑟𝑝𝑟𝑒𝑡 𝑡ℎ𝑒 𝑐𝑜𝑒𝑓𝑓𝑖𝑐𝑖𝑒𝑛𝑡

(3) Correlation

→ 𝐼𝑚𝑝𝑜𝑟𝑡𝑎𝑛𝑡 𝑡𝑜 𝑘𝑛𝑜𝑤 𝑡ℎ𝑒 𝑟𝑒𝑙𝑎𝑡𝑖𝑜𝑛𝑠ℎ𝑖𝑝 𝑏𝑒𝑡𝑤𝑒𝑒𝑛 𝑡ℎ𝑒 𝑡𝑤𝑜 𝑣𝑎𝑟𝑖𝑎𝑏𝑙𝑒𝑠


𝐼𝑡 𝑎𝑙𝑠𝑜 𝑖𝑛𝑑𝑖𝑐𝑎𝑡𝑒 𝑡ℎ𝑒 𝑠𝑡𝑟𝑒𝑛𝑔𝑡ℎ 𝑜𝑓 𝑡ℎ𝑖𝑠 𝑟𝑒𝑙𝑎𝑡𝑖𝑜𝑛𝑠ℎ𝑖𝑝

• Given that Correlation is comprehensive, why do we compute covariance?

o In Correlation we forced a given weight 𝜎𝑋 . 𝜎𝑌 to make it between [-1,+1].


o If we want to derive a theory in statistics, we should not depart from given
weights but from the formulas themselves but without any weights.
o When we derive a theory or develop further analysis in statistics we start
from covariance (we should not derive a theory based on some weights).
Two Variables can be dependent on each other but their Covariance is Zero!

Elaboration:
We assume that the Variable Y is dependent on the variable X
𝑌 = 𝑍𝑋
𝑋 𝑍 𝑌 = 𝑍𝑋
6.062178 1 6.062178
0.866025 1 0.866025
6.062178 -1 -6.062178
0.866025 -1 -0.866025
Cov(X,Y) 0

We can conclude that Covariance and Correlation capture the linear relationship
between the two variables. Non-linear or more complex relationship are not
captured by those two indicators.

You might also like