You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Copy file name to clipboardExpand all lines: lectures/cross_section.md
+85-89Lines changed: 85 additions & 89 deletions
Original file line number
Diff line number
Diff line change
@@ -518,95 +518,6 @@ Replicate {ref}`the figure presented above <light_heavy_fig1>` that compares nor
518
518
Use `np.random.seed(11)` to set the seed.
519
519
```
520
520
521
-
522
-
```{exercise}
523
-
:label: ht_ex2
524
-
525
-
Prove: If $X$ has a Pareto tail with tail index $\alpha$, then
526
-
$\mathbb E[X^r] = \infty$ for all $r \geq \alpha$.
527
-
```
528
-
529
-
530
-
```{exercise}
531
-
:label: ht_ex3
532
-
533
-
Repeat exercise 1, but replace the three distributions (two normal, one
534
-
Cauchy) with three Pareto distributions using different choices of
535
-
$\alpha$.
536
-
537
-
For $\alpha$, try 1.15, 1.5 and 1.75.
538
-
539
-
Use `np.random.seed(11)` to set the seed.
540
-
```
541
-
542
-
543
-
```{exercise}
544
-
:label: ht_ex4
545
-
546
-
547
-
Replicate the rank-size plot figure {ref}`presented above <rank_size_fig1>`.
548
-
549
-
If you like you can use the function `qe.rank_size` from the `quantecon` library to generate the plots.
550
-
551
-
Use `np.random.seed(13)` to set the seed.
552
-
```
553
-
554
-
```{exercise}
555
-
:label: ht_ex5
556
-
557
-
There is an ongoing argument about whether the firm size distribution should
558
-
be modeled as a Pareto distribution or a lognormal distribution (see, e.g.,
559
-
{cite}`fujiwara2004pareto`, {cite}`kondo2018us` or {cite}`schluter2019size`).
560
-
561
-
This sounds esoteric but has real implications for a variety of economic
562
-
phenomena.
563
-
564
-
To illustrate this fact in a simple way, let us consider an economy with
565
-
100,000 firms, an interest rate of `r = 0.05` and a corporate tax rate of
566
-
15%.
567
-
568
-
Your task is to estimate the present discounted value of projected corporate
569
-
tax revenue over the next 10 years.
570
-
571
-
Because we are forecasting, we need a model.
572
-
573
-
We will suppose that
574
-
575
-
1. the number of firms and the firm size distribution (measured in profits) remain fixed and
576
-
1. the firm size distribution is either lognormal or Pareto.
577
-
578
-
Present discounted value of tax revenue will be estimated by
579
-
580
-
1. generating 100,000 draws of firm profit from the firm size distribution,
581
-
1. multiplying by the tax rate, and
582
-
1. summing the results with discounting to obtain present value.
583
-
584
-
If $X$ has the Pareto distribution, then there are positive constants $\bar x$
585
-
and $\alpha$ such that
586
-
587
-
The Pareto distribution is assumed to take the form {eq}`pareto` with $\bar x = 1$ and $\alpha = 1.05$.
588
-
589
-
(The value the tail index $\alpha$ is plausible given the data {cite}`gabaix2016power`.)
590
-
591
-
To make the lognormal option as similar as possible to the Pareto option, choose its parameters such that the mean and median of both distributions are the same.
592
-
593
-
Note that, for each distribution, your estimate of tax revenue will be random because it is based on a finite number of draws.
594
-
595
-
To take this into account, generate 100 replications (evaluations of tax revenue) for each of the two distributions and compare the two samples by
596
-
597
-
* producing a [violin plot](https://en.wikipedia.org/wiki/Violin_plot) visualizing the two samples side-by-side and
598
-
* printing the mean and standard deviation of both samples.
599
-
600
-
For the seed use `np.random.seed(1234)`.
601
-
602
-
What differences do you observe?
603
-
604
-
(Note: a better approach to this problem would be to model firm dynamics and
605
-
try to track individual firms given the current distribution. We will discuss
606
-
firm dynamics in later lectures.)
607
-
```
608
-
609
-
610
521
```{solution-start} ht_ex1
611
522
:class: dropdown
612
523
```
@@ -644,6 +555,13 @@ plt.show()
644
555
```
645
556
646
557
558
+
```{exercise}
559
+
:label: ht_ex2
560
+
561
+
Prove: If $X$ has a Pareto tail with tail index $\alpha$, then
562
+
$\mathbb E[X^r] = \infty$ for all $r \geq \alpha$.
563
+
```
564
+
647
565
```{solution-start} ht_ex2
648
566
:class: dropdown
649
567
```
@@ -674,6 +592,20 @@ Since $r \geq \alpha$, we have $\mathbb E X^r = \infty$.
674
592
```{solution-end}
675
593
```
676
594
595
+
596
+
```{exercise}
597
+
:label: ht_ex3
598
+
599
+
Repeat exercise 1, but replace the three distributions (two normal, one
600
+
Cauchy) with three Pareto distributions using different choices of
601
+
$\alpha$.
602
+
603
+
For $\alpha$, try 1.15, 1.5 and 1.75.
604
+
605
+
Use `np.random.seed(11)` to set the seed.
606
+
```
607
+
608
+
677
609
```{solution-start} ht_ex3
678
610
:class: dropdown
679
611
```
@@ -704,6 +636,17 @@ plt.show()
704
636
```
705
637
706
638
639
+
```{exercise}
640
+
:label: ht_ex4
641
+
642
+
643
+
Replicate the rank-size plot figure {ref}`presented above <rank_size_fig1>`.
644
+
645
+
If you like you can use the function `qe.rank_size` from the `quantecon` library to generate the plots.
646
+
647
+
Use `np.random.seed(13)` to set the seed.
648
+
```
649
+
707
650
```{solution-start} ht_ex4
708
651
:class: dropdown
709
652
```
@@ -747,7 +690,60 @@ plt.show()
747
690
```{solution-end}
748
691
```
749
692
693
+
```{exercise}
694
+
:label: ht_ex5
750
695
696
+
There is an ongoing argument about whether the firm size distribution should
697
+
be modeled as a Pareto distribution or a lognormal distribution (see, e.g.,
698
+
{cite}`fujiwara2004pareto`, {cite}`kondo2018us` or {cite}`schluter2019size`).
699
+
700
+
This sounds esoteric but has real implications for a variety of economic
701
+
phenomena.
702
+
703
+
To illustrate this fact in a simple way, let us consider an economy with
704
+
100,000 firms, an interest rate of `r = 0.05` and a corporate tax rate of
705
+
15%.
706
+
707
+
Your task is to estimate the present discounted value of projected corporate
708
+
tax revenue over the next 10 years.
709
+
710
+
Because we are forecasting, we need a model.
711
+
712
+
We will suppose that
713
+
714
+
1. the number of firms and the firm size distribution (measured in profits) remain fixed and
715
+
1. the firm size distribution is either lognormal or Pareto.
716
+
717
+
Present discounted value of tax revenue will be estimated by
718
+
719
+
1. generating 100,000 draws of firm profit from the firm size distribution,
720
+
1. multiplying by the tax rate, and
721
+
1. summing the results with discounting to obtain present value.
722
+
723
+
If $X$ has the Pareto distribution, then there are positive constants $\bar x$
724
+
and $\alpha$ such that
725
+
726
+
The Pareto distribution is assumed to take the form {eq}`pareto` with $\bar x = 1$ and $\alpha = 1.05$.
727
+
728
+
(The value the tail index $\alpha$ is plausible given the data {cite}`gabaix2016power`.)
729
+
730
+
To make the lognormal option as similar as possible to the Pareto option, choose its parameters such that the mean and median of both distributions are the same.
731
+
732
+
Note that, for each distribution, your estimate of tax revenue will be random because it is based on a finite number of draws.
733
+
734
+
To take this into account, generate 100 replications (evaluations of tax revenue) for each of the two distributions and compare the two samples by
735
+
736
+
* producing a [violin plot](https://en.wikipedia.org/wiki/Violin_plot) visualizing the two samples side-by-side and
737
+
* printing the mean and standard deviation of both samples.
738
+
739
+
For the seed use `np.random.seed(1234)`.
740
+
741
+
What differences do you observe?
742
+
743
+
(Note: a better approach to this problem would be to model firm dynamics and
744
+
try to track individual firms given the current distribution. We will discuss
0 commit comments