S&P 500: Great Depression vs. 2008 Financial Crisis
William Chiu
5/17/2020
Summary
The stock market recovery from the Great Depression (of 1929) was very slow compared to the recovery from the Financial Crisis (of 2008). The document compares the daily performance of $10,000 invested one month prior to the beginning of each recession.
A portfolio of $10,000 at the beginning of the Great Depression never recovers even after 12.5 years. In stark contrast, a portfolio of $10,000 recovers approximately 5.3 years after the Financial Crisis.
I thank the author of this R tutorial. I also want to thank the author of this other tutorial. And this tutorial. One of the best things about using R is the ease of copying and pasting other people’s code.
Advanced users may adjust the ticker symbol and initial investment dates in the chunk below and re-run the analysis.
### document parameters
ticker <- "^GSPC"
GDstartdate <- '1929-07-01'
FCstartdate <- '2007-11-01'Data
Extract historical data for the S&P 500 from Yahoo Finance.
library(tidyverse)
library(tidyquant)
GSPC <- tq_get(ticker, from = '1927-01-01',
to = "2020-05-17",
get = "stock.prices")
head(GSPC)## # A tibble: 6 x 8
## symbol date open high low close volume adjusted
## <chr> <date> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl>
## 1 ^GSPC 1927-12-30 17.7 17.7 17.7 17.7 0 17.7
## 2 ^GSPC 1928-01-03 17.8 17.8 17.8 17.8 0 17.8
## 3 ^GSPC 1928-01-04 17.7 17.7 17.7 17.7 0 17.7
## 4 ^GSPC 1928-01-05 17.5 17.5 17.5 17.5 0 17.5
## 5 ^GSPC 1928-01-06 17.7 17.7 17.7 17.7 0 17.7
## 6 ^GSPC 1928-01-09 17.5 17.5 17.5 17.5 0 17.5Recession Dates
According to the NBER the Great Depression began on August 1929 while the 2008 Financial Crisis began on December 2007. In order to simulate the performance of a portfolio that began immediately prior to the beginning of the recession, the document assumes $10,000 was invested in the S&P 500 on the following dates:
- 1929-07-01
- 2007-11-01
GSPC %>%
filter(date >= GDstartdate) %>%
head()## # A tibble: 6 x 8
## symbol date open high low close volume adjusted
## <chr> <date> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl>
## 1 ^GSPC 1929-07-01 27.8 27.8 27.8 27.8 0 27.8
## 2 ^GSPC 1929-07-02 28.1 28.1 28.1 28.1 0 28.1
## 3 ^GSPC 1929-07-03 28.2 28.2 28.2 28.2 0 28.2
## 4 ^GSPC 1929-07-05 28.3 28.3 28.3 28.3 0 28.3
## 5 ^GSPC 1929-07-08 28.4 28.4 28.4 28.4 0 28.4
## 6 ^GSPC 1929-07-09 28.3 28.3 28.3 28.3 0 28.3GSPC %>%
filter (date >= FCstartdate) %>%
head()## # A tibble: 6 x 8
## symbol date open high low close volume adjusted
## <chr> <date> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl>
## 1 ^GSPC 2007-11-01 1546. 1546. 1507. 1508. 4241470000 1508.
## 2 ^GSPC 2007-11-02 1511. 1513. 1493. 1510. 4285990000 1510.
## 3 ^GSPC 2007-11-05 1506. 1511. 1490. 1502. 3819330000 1502.
## 4 ^GSPC 2007-11-06 1505. 1521. 1499. 1520. 3879160000 1520.
## 5 ^GSPC 2007-11-07 1515. 1515. 1475. 1476. 4353160000 1476.
## 6 ^GSPC 2007-11-08 1475. 1482. 1450. 1475. 5439720000 1475.Splitting the Data
Split the data sets between Great Depression and Financial Crisis. Then add new variable for portfolio age (i.e., the number of days since the initial $10,000 investment). Use calendar days rather than trading days.
GD <- GSPC %>%
filter(date >= GDstartdate) %>%
mutate(PortfolioAge = interval(first(date), date) / days(1))
FC <- GSPC %>%
filter (date >= FCstartdate) %>%
mutate(PortfolioAge = interval(first(date), date) / days(1))
head(GD)## # A tibble: 6 x 9
## symbol date open high low close volume adjusted PortfolioAge
## <chr> <date> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl>
## 1 ^GSPC 1929-07-01 27.8 27.8 27.8 27.8 0 27.8 0
## 2 ^GSPC 1929-07-02 28.1 28.1 28.1 28.1 0 28.1 1
## 3 ^GSPC 1929-07-03 28.2 28.2 28.2 28.2 0 28.2 2
## 4 ^GSPC 1929-07-05 28.3 28.3 28.3 28.3 0 28.3 4
## 5 ^GSPC 1929-07-08 28.4 28.4 28.4 28.4 0 28.4 7
## 6 ^GSPC 1929-07-09 28.3 28.3 28.3 28.3 0 28.3 8head(FC)## # A tibble: 6 x 9
## symbol date open high low close volume adjusted PortfolioAge
## <chr> <date> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl>
## 1 ^GSPC 2007-11-01 1546. 1546. 1507. 1508. 4241470000 1508. 0
## 2 ^GSPC 2007-11-02 1511. 1513. 1493. 1510. 4285990000 1510. 1
## 3 ^GSPC 2007-11-05 1506. 1511. 1490. 1502. 3819330000 1502. 4
## 4 ^GSPC 2007-11-06 1505. 1521. 1499. 1520. 3879160000 1520. 5
## 5 ^GSPC 2007-11-07 1515. 1515. 1475. 1476. 4353160000 1476. 6
## 6 ^GSPC 2007-11-08 1475. 1482. 1450. 1475. 5439720000 1475. 7Calculate the cumulative returns
GD2 <- GD %>%
mutate(CumReturn1929 = adjusted / first(adjusted)) %>%
select(PortfolioAge, date, adjusted, CumReturn1929)
FC2 <- FC %>%
mutate(CumReturn2008 = adjusted / first(adjusted)) %>%
select(PortfolioAge, date, adjusted, CumReturn2008)
head(GD2)## # A tibble: 6 x 4
## PortfolioAge date adjusted CumReturn1929
## <dbl> <date> <dbl> <dbl>
## 1 0 1929-07-01 27.8 1
## 2 1 1929-07-02 28.1 1.01
## 3 2 1929-07-03 28.2 1.01
## 4 4 1929-07-05 28.3 1.02
## 5 7 1929-07-08 28.4 1.02
## 6 8 1929-07-09 28.3 1.02head(FC2)## # A tibble: 6 x 4
## PortfolioAge date adjusted CumReturn2008
## <dbl> <date> <dbl> <dbl>
## 1 0 2007-11-01 1508. 1
## 2 1 2007-11-02 1510. 1.00
## 3 4 2007-11-05 1502. 0.996
## 4 5 2007-11-06 1520. 1.01
## 5 6 2007-11-07 1476. 0.978
## 6 7 2007-11-08 1475. 0.978Merge the returns by Portfolio Age
GD_FC <- GD2 %>%
inner_join(FC2, by="PortfolioAge") %>%
select(PortfolioAge, CumReturn1929, CumReturn2008)
head(GD_FC)## # A tibble: 6 x 3
## PortfolioAge CumReturn1929 CumReturn2008
## <dbl> <dbl> <dbl>
## 1 0 1 1
## 2 1 1.01 1.00
## 3 4 1.02 0.996
## 4 7 1.02 0.978
## 5 8 1.02 0.964
## 6 11 1.03 0.954We’re almost done. We need to multiply the cumulative returns by $10,000 to get the value of the portfolio in dollar terms.
GD_FC2 <- GD_FC %>%
mutate(GreatDepression1929 = CumReturn1929*10000,
FinancialCrisis2008 = CumReturn2008*10000)
head(GD_FC2)## # A tibble: 6 x 5
## PortfolioAge CumReturn1929 CumReturn2008 GreatDepression19… FinancialCrisis20…
## <dbl> <dbl> <dbl> <dbl> <dbl>
## 1 0 1 1 10000 10000
## 2 1 1.01 1.00 10104. 10008.
## 3 4 1.02 0.996 10198. 9958.
## 4 7 1.02 0.978 10216. 9777.
## 5 8 1.02 0.964 10184. 9637.
## 6 11 1.03 0.954 10288. 9541.Convert columns to rows
This part is a little tricky. The columns GreatDepression1929 and FinancialCrisis2008 should be converted into rows with a new column (i.e., Type) to distinguish between the 1929 and 2008 portfolios. I perform this step to make plotting easier.
GD_FC2_gather <- GD_FC2 %>%
select(PortfolioAge, GreatDepression1929, FinancialCrisis2008) %>%
gather("Type", "PortfolioValue", -PortfolioAge)
head(GD_FC2_gather)## # A tibble: 6 x 3
## PortfolioAge Type PortfolioValue
## <dbl> <chr> <dbl>
## 1 0 GreatDepression1929 10000
## 2 1 GreatDepression1929 10104.
## 3 4 GreatDepression1929 10198.
## 4 7 GreatDepression1929 10216.
## 5 8 GreatDepression1929 10184.
## 6 11 GreatDepression1929 10288.tail(GD_FC2_gather)## # A tibble: 6 x 3
## PortfolioAge Type PortfolioValue
## <dbl> <chr> <dbl>
## 1 4568 FinancialCrisis2008 18846.
## 2 4571 FinancialCrisis2008 19100.
## 3 4572 FinancialCrisis2008 19423.
## 4 4575 FinancialCrisis2008 19425.
## 5 4578 FinancialCrisis2008 18910.
## 6 4579 FinancialCrisis2008 18985.Plot
ggplot(GD_FC2_gather, aes(x=PortfolioAge, y=PortfolioValue)) +
geom_line(aes(color=Type)) +
theme_classic() +
geom_hline(yintercept=10000, linetype="dashed", color = "black") +
scale_y_continuous(breaks=seq(0,25000,2500)) +
scale_x_continuous(name="Portfolio Age (in days)",breaks=seq(0,365*13,365)) +
scale_color_manual(values=c('blue','orange')) +
ggtitle("Investing $10,000 in the S&P 500")
Conclusion
Had you invested $10,000 on 1929-07-01, your portfolio would have been worth about $3,000 after 12.5 years. In stark contrast, if you invested $10,000 on 2007-11-01, your portfolio would have been worth about $19,000 after 12.5 years.