Mulbrandon Research, Design & Insights

Here is a simple graph comparing the variation of annual returns for US stocks, US 10-year bonds and US 3-month t-bills. I have included both nominal returns (not adjusted for inflation) and real returns. 

Stocks are of course the most risky of the three with both the highest and the lowest  returns then comes 10-year bonds and finally t-bills with the smallest but the most consistant returns. However, after 3-month t-bills are adjusted for inflation there are many years you will "lose" money. And in years with deflation, your real return will be larger than the nominal return (i.e. you did better holding stocks in 1933 when you take into account the effects of deflation in that year while 1954 had the highest nominal stock return).

Return Data from Damodaran Online | Updated Data | Historical Returns on Stock Bonds and Bills - Untied States CPI form Measuring Worth. Created using OnmiGraphSketcher.

I dug up at the US Census Bureau serveral reports about family and individual income and created a series of graphs plotting the income distribution of households under $100,000 a year adjusted for inflation. (Pages 17, 18, 19 from my Income Guide)

I am defining middle-income households as $30,000-$80,000. One of the stories these graph tell is that for 20+ years after 1945 more households entered the "middle class". However, over the next 40 years, the percent of middle-income households shrank in part because the percent of households with more than $80,000 a year grew. 

Graphs created in OmniGraphSketcher and annotated in Illustrator. Data from the US Census. You take a look at some of the older reports they have online here:

US Census Bureau. “Families and Individual Money Income in the United States: 1945. Table 2.” September 2011. http://www2.census.gov/prod2/popscan/p60-002.pdf.

———. “Income of Families and Persons in the United States: 1950. Table 1.” September 2011.  http://www2.census.gov/prod2/popscan/p60-009.pdf.

———. “Income of Families and Persons in the United States: 1960. Table 5.” September 2011.  http://www2.census.gov/prod2/popscan/p60-037.pdf.

———. “Income, Poverty, and Health Insurance Coverage in the United States: 2010. Table A-2.” September 2011. http://www.census.gov/prod/2011pubs/p60-239.pdf.

Copies of this graphic can be purchased at Zazzle

Inflation is when prices go up. However, inflation's relationship to stock prices can be a little more complex. In this graph, I am revisiting historical data I used several years ago in a series of historical graphs looking at the stock market (see GDP per Capita vs US Stock Prices and Real Growth in Stock Returns Dividends Reinvested).

I graphed the price/earnings ratio back to 1880 and highlighted the years that inflation was high and when there was deflation. The P/E is calculated from trailing 10-year earnings. The low inflation years (5% or less) had the highest P/E while the high inflation years has the lowest P/Es. Deflation years had P/Es near the average.

S&P data from Robert Shiller and CPI data from MeasuringWorth. Graphic was created using OmniGraphSketcher.

Let me begin with a disclaimer. This industry includes firms that are pure internet-based activities, like hosting and web searches as well as ones that are being disrupted by the internet like newspapers and broadcasting.

Information Industry's share of GDP 1947-2010 

Starting with the Information industry's share of GDP (from page 119 of my book), while it has grown over the last 50 years it is still around 4% of GDP as of 2010. 

However, the Information industry represents only 2% of the 170 million jobs in the US economy. 

In this treemap of all occupations in the US economy, each occupation is represented by a rectangle, the bigger it is the more jobs it has. Look at the upper right corner to see Information's share of jobs. The dark red represents the percentage of jobs loss and the only area more red than the Information sector is Manufacturing.

Now drilling into just the Information industry, some of the bigger occupations are: editors, computer software engineers, customer service representatives, telecommunications equipment installers, reporters and finally producers & directors. Out of that list only producers & directors had job growth between 2001-2011. 

If you are interested in the income of these occupations or want to explore additional industries take a look at my book. 

Data used in these graphics was based on BLS Occupational Handbook (provided by 
EMSI) and value-added GDP from the BEA. The area graph was created in Omnigraphsketcher and the treemaps by R. Label were later added using Illustrator.

From a very interesting database of Texas Government Employee Salaries run by The Texas Tribune, I created three data graphics for An Illustrated Guide of Income in the United States showing their distribution (pages 92–94).

I start with graphing the distribution of all public employee salaries below $250,000, 99.7% of all state employees, listing the most common job titles I found: teachers and professors, police officers, clerks and administrative assistants, bus operators, child protective services specialist and mental retardation assistants.

However, to graph the long tail of the income distribution I have to graph individual salaries. Many are heads of surgery departments and head coaches. The highest salary goes to the head football coach at the University of Texas at Austin. (Salary of $2.5 million plus a bonuses of $2.7 million for a total of $5.2 million)

Another way to show how much inequality exists with these salaries is to plot the cumulative share of Texas public employees vs cumulative share of their salaries. But how does this compare to the Untied States as a whole? I found in a report from the CBO, a graph plotting the 2007 cumulative share of all US households against household income which I used as a stand in for everyone in the US. You can see that the US household income distribution is more unequal than salaries of Texas state employees.

Data sources: The Texas Tribune and the Congressional Budget Office.

Design notes: Graphs were created using OmniGraphSketcher, copied into Adobe Illustrator where annotations were added. The illustrator file was then placed into an InDesign document for the book. View all the graphics from the book online.

One of the most popular maps I have on my site I got from a site called Social Explorer. It is about the 2000 poverty rates across the US. For my Illustrated Guide to Income in the United States, I created an new map using data through 2010 and a "divergent" color scheme which sets the light gray midpoint at the national poverty rate of 14%. Red are counties with a higher rate of poverty than the national rate while blue counties have a lower rate.

I also looked at the households with more than $200,000 a year and mapped where they are clustered. (BTW, if you want to see the voting patterns of high-income counties  and find out if the wealthy counties voted for Obama or Romney check out a map I created for Design & Geography). 

Data is from the US Census Bureau, American Community Survey.

First, what do I mean by good job? Well-paid? High level of job satisfaction? Lots of job security?

From heteconomist: Why so many jobs are crappy

Satisfying jobs – let's call them 'good jobs' – will generally be ones where learning occurs at a steady pace more or less indefinitely, probably as part of a defined career path. [...] Once you gain experience in a good job, you will soon become much more efficient in the role than an inexperienced replacement would be.

Summarizing this post, MarginalRevolution comments

The first key point is that if you learn more on the job on a regular basis (i.e., your job is interesting), you become harder to replace from the point of view of your boss.  Over time you win more of the bargaining surplus. 

So looking at my treemaps of occupations from An Illustrated Guide to Income in the United States, I have a category called "Long-term on-the-job training" defined as:

More than 12 months of on-the-job training or, alternatively, combined work experience and formal classroom instruction, are needed for workers to develop the skills to attain competency. Training is occupation specific rather than job specific; therefore, skills learned can be transferred to another job in the same occupation. This on-the-job training category also includes employer-sponsored training programs.

This sound like some of the "good jobs" people are looking for that don't require a college degree.

Click on images to enlarge.

WIth this data graphic, I wanted to give an overview of all occupations grouped by the education/training required. The size of each rectangle represents the number of jobs.

Occupations like: electricians, carpenters, farmers & ranchers, restaurant cooks, photographers, police & sheriff's patrol offices can be found in  "Long-term on-the-job training".

The last three occupations (restaurants cooks, photographers, police & sheriff's patrol offices) increased in number of jobs from 2001-2011.

To learn more about which industries have a high number of  "long-term on-the-job training" jobs, first check out pages 128-148 in my book and then goto the BLS's Occupational Handbook to search for more jobs with long-term on-the-job training but no college requirement.

Data for the occupation treemaps was provide by EMSI based on data from Bureau of Labor Statistics (BLS).
.

More excerpts from An Illustrated Guide to Income in the United States (Pages 66, 68, 69).

In my previous post, I looked at wages, showing that for many occupations, they have not kept up with the overall growth of the economy. But what about the growing number households with incomes well above the national average of $67,500? (Focus on the orange lines/area in the 3 graphs below)

Click on the images for a closer look. This will open a lightbox the same size as the browser window.

In 1945, less than 10% of family households had incomes above $80,000 (adjusted for inflation) steadily increasing to 30% in the late 1990s. They continued to increase up until 2000 even while wages flatten or dropped.

Looking at the bottom graph, the number of family households with two earners grew and surpassed number of single earner household in the late 1960s. This growth mirrors the growth of families with $80,000 or more a year. One can easily assume that second earner in these households were women as women entered the labor force in larger numbers boosting their family's income. 

Note: Single people living alone are not included in "Family Households," bottom graph, but are included in "All Households," top graph.

In 2010, 72% of the households with income above $80,000 a year have two or more earners compared to 41% of households with income between $30,000 and $80,000. It is this increase in the number of two-earner households that accounts for household incomes increasing when wages have not.

To learn more about Income in the US buy or read the book.

Data for households came from the US Census

US Census Bureau. “Historical Income Tables: Households.” June 2011. http://www.census.gov/hhes/www/income/data/historical/household/index.html.

US Census Bureau. “Table HINC-01. Selected Characteristics of Households, by Total Money Income in 2010.” 2012. http://www.census.gov/hhes/www/cpstables/032011/hhinc/new01_001.htm

 (See bibliography for additional data)

Below are five data graphics from my new book An Illustrated Guide to Income in the United States (pgs 106, 108, 109, 110, 112) that shows the long-term growth in wages in the US.

Click on the images for a closer look. This will open a lightbox the same size as the browser window.

Over the last couple of centuries there has been a steady increase in wages for both unskilled workers...

....and production workers. A lot of this growth is a result of the increases in worker productivity due the industrial revolution of the late 1770s and 1800s. However, over the last 40 years, this long-term growth has stopped or slowed down...

Updated 3/25/13: This Production Workers series includes benefits all the way back to when benefits became measurable in the early 1900s. More detailed definition can be found here.

...even though the GDP per person continues to grow. At the same time, the growth rate of GDP per worker has slowed compared to the overall growth of the economy.

Looking at just goods-producing industries, wages dropped for manufacturing, construction, and mining & logging since their a peak in the 1970s.

But among so-called service industries over the same time period there has been either a dip in wages or no real growth. Exceptions include jobs in the financial industry, education & health services and "other" services (which is mashup of occupations like auto mechanics, pet care, promoting political causes or religious activities).

Data Sources for Wages (See bibliography for more references)

Officer, Lawrence H., and Samuel H. Williamson. “Annual Wages in the United States, 1774–Present.” MeasuringWorth.com, 2011. http://www.measuringworth.com/uswage/.

US Bureau of Labor Statistics. “Table B-8. Average hourly and weekly earnings of production and nonsupervisory employees on private nonfarm payrolls by industry sector, seasonally adjusted.” November 2012. http://www.bls.gov/webapps/legacy/cesbtab8.htm.

Designer's Notes: Some of my thoughts on the design and the approaches I used. 

One of the many data series that have exponential growth. They are a challenge  since I want to make the graph accessible to a wide (non-finance) audience but log graphs are very helpful when you need to see the relative changes across the entire time series.

Through out the book you will see me stacking two or more graphs when they have same timeline on the x-axies. EIther I want to facility comparisons of different time series or in this case show the same data on a "normal" y-axis and a log scale. You may also notice the absence of the GDP per person on this graph. SInce I was working with average hourly wages over 200-year period I didn't feel comfortable converting it to an annual salary for production workers who can often work irregular hours.

Unlike the previous two pages, values of the salaries was not the focus of this graphic. In this case I wanted to compare the change over time of GDP per person vs GDP per worker so I converted the series so they both started from the same point in 1947 (the first year I had the data for the civilian labor force). 

When looking at the historical wages for the major industries, I tried several ways to group the series hoping I could show some interesting patterns. I finally went with the simplest approach and created one graph for goods-producing industries... 

...and a second graph for service-providing industries. While this worked well for the first graph, 3 time series with similar behavior, with the service graph I was left with a less optimal design, a "spaghetti graph", which I normally try to avoid. I decided to keep scale the same on both graphs to show the overall pattern for the goods and service industries rather than splitting each series into separate smaller line charts.  

In each case, the original graph was created in OmniGraphSketcher. Additional annotations were added in Illustrator and when multiple graphs are need they are laid out in Illustrator. Finally, I link to the Illustrator file from within an InDesign document where i add page titles, a footer for data sources and pagination.

I wanted to let you that after 2 years of work I have just published An Illustrated Guide to Income in the United States. You can peek inside the book here. 

Over the next month, I will be posting highlights from the book and describing some of the challenges of making it.  Enjoy!  Catherine

Recently, I gave a talk at a design lecture called Creative Mornings. I only had 8 mins so I decided to present one of my infographics from my upcoming book An Illustrated Guide to Income in the United States that uses treemaps to visualize the distribution of income (1975 vs 2008). My part of the talk begins at 9:34.

An infographic I worked on for the NYC Comptroller report on eduction called Beyond High School. I created this in Adobe Illustrator.

I was on a data visualization panel early this month at the Association for Public Policy Analysis and Management annual meeting (with Kurt Voelker, Bryan Connor and Jon Schwabish who organized the panel. It was great to see everyone turn out for what I believe was their first dataviz panel. Here are the slides I presented which are similar to the presentation I gave American Association for Budget and Program Analysis 2012 Spring Symposium but with some edits. 

I was asked a while back to work on a cover for a report “Income Inequality in New York City” published by New York City Comptroller’s Office. Their analysis of the 2009 New York City Income Tax Files breakout the share of income going to the top 1% and other income groups. While my graphics were not used for the cover, you can take a look at two versions I mocked up based on some their data and cover ideas.

I had a chance to participate on a panel about Visualizing Data at the American Association for Budget and Program Analysis 2012 Spring Symposium (lots of talks for budget geeks) along with Jonathan Schwabish from Congressional Budget Office and Ellie Fields from Tableau Software. We discuss the analysis, design, and editing process for creating data graphics. If you made it to the session and would like to see my slides again or missed it but want to learn about the methods and resources I find most helpful take a look:

Here is a sneak peek at my An Illustrated Guide to Income in the United States. These are a set of data graphics looking at the average income and change in number of jobs over the last ten years for 800+ occupation by industry and by education. Be sure to sign up to be notified when the Income Guide is done. 

Data from EMSI

Due to popular demand, I have updated my 2010 graph on top marginal tax rates. In addition, during this year's tax season, I will be selling copies of my Top Marginal Tax Rates graph as a tabloid size 11"x17" poster.  

FYI, your marginal tax rate is the rate you pay on the "last dollar" you earn; but when you view the taxes you paid as a percentage of your income, your effective tax rate is less than your marginal rate, especially after you take into account the deductions and exemptions, i.e. income that is not subject to any tax.

Tax Data: Married filing jointly, Capital Gains & Regular,Historical Corporate, Corporate Tax Schedule (page 16) pdf

First published in Slate to accompany an article written by Tim Noah, I created these graphs about income inequality covering the changes in income inequality as well as looking at changes in race, gender, education, taxes and political party in the White House. 

I worked with Economic Modeling Specialists on a data graphic that looks to distinguish between growth from large national forces vs. local competitive advantages within a state. Learn more about the data and analysis at the EMSI Blog.

This one of the graphics that I presented recently at The Big Picture conference here in New York City. It is from a project I am currently working on called An Illustrated Guide to Income in the United States: a collection of infographics, maps and charts looking at the different incomes and occupations in the United States.

Recently the conversation in the news has been about the top 1%, however, in this graphic I show the breakdown of personal income by different percentiles, including the top 0.01% (i.e. income above $9 million). I have used 10,000 "people" to represent the tax returns filed in 2008, each "person" one equals 15,246 tax units. (A tax unit is single adult or married couple living together, including their dependents.)

So the top 1% are represented by the 100 "people" in the four (orange, yellow, magenta & red) rectangles the upper left corner.

Approximately $8.2 trillion in personal income (including capital gains) was reported to the IRS in 2008. Divide that by 152 million tax units you get an average income of $54,315. I have the size of the "people" represent the average income for each percentile group. For example the Average Income for the Top 0.01% = $27 million.

Data is from Saez and Piketty research which is now available at the The World Top Incomes Database 

Sign up to be notified when the Income Guide is done:

Enter your email address:  Powered by TinyLetter