Thank you! The charts really bothered me and this shows that the article is portraying the trends a little more aggressively than they really are.
I still think having only two data points (one during the Great Recession and one in 2018) reeks of cherry picking. I’d like the author to have provided more historical data to demonstrate the effect.
Agree, it's not a very good article but if you're a regular Economist reader you know that they frequently put out charts like this exactly to highlight interesting but potentially misleading correlations.
Click on the "Daily Chart" link in the header to see more examples.
"A smaller relative slice of a much, much larger pie can still be a much larger piece of pie in absolute terms. The poor are not getting poorer, the rich are just getting richer faster."
I don't see where the article claims otherwise.
That said, the point is pedantic: if the share of income in a group gets smaller over time, their purchasing power goes down correspondingly as well. The overall size of the pie matters only to economists.
"Also, the choice of timeline is very suspicious. Pegging the start of the graphs at the end of the Great Depression / WWII skews the graph, as does omitting the context of the 19th and early 20th centuries. This is more about chart-junk supporting a pre-conceived narrative than anything else."
And not including the civil war skews the graph as well. Or the depression of 1893. You can always find an argument like that to justify not believing something that you choose not to believe. Calling it "chart junk" is just exaggeration.
That’s one of the problems with data. It seems so exact but can really be so misleading. You break out charts, I break out charts. [1] My gut feeling is that much of middle class america has seen significant wage and livelihood stagnation. I’m all for free trade and a competitive economy, but something feels off about the past couple decades and the distribution of economic gains. And I don’t really trust the Federal Reserve and their CPI metrics—the institution is far too political and the members have too much skin in the game, rotating in and out of banking jobs.
I stopped reading at the date ranges. It’s comparing the decade 1983-93 (start at recovery from long recession, measure across famous economic boom with only short 1990 downturn) versus 2003-2013. (Oldish data anyway and measures across the famously deep 2008 recession which had an unusually slow recovery.) and you want to make a point about economic mobility and inequality in 2019 with this? We may have problems with economic mobility but this is a very bad version of that argument.
1. 2022 was right in the middle of the disruption, which is now over.
2. You're showing a derivative chart and not the value, which amplifies noise into signal. What if it showed a giant leap in the following quarter (which it does!), would you change your mind from "it's terrible" to "it's amazing"?
3. Disposable income tends to be a noisy statistic anyway as it's a delta statistic conflates motion of two different measurables (income and expenses) that are perfectly measurable by themselves.
4. You're just wrong anyway. Here's the FRED chart (always go to FRED, always, everyone else is peddling something) of exactly the statistic you posted. Tell me if this chart is consistent with the story you interpreted:
The chart title comes from the actual paper (linked right above the graph: https://www.nber.org/system/files/working_papers/w28912/w289...) that includes two charts (COVID-19 and Great Recession), but the article only includes the first chart (so it makes less sense in the context).
Though I find his point intriguing about corporate America entrenching itself, the graphs he has chosen to represent it are a bit misrepresented - most notably, the graph about startup failure rate. I disagree that the rate of failure in startups has risen significantly in the last 20 years - he cherry picked his data point from 1991 to 2011. The first fifteen years of that interval remain almost entirely flat. The spike in the last three are easily explainable as part of the housing bubble bursting that sent a bunch of businesses in the US to the deadpool. Conversely, the declines in new businesses in the early 80s and late 00s can be easily attributed to spikes in the price of oil and the housing bubble bursting. The rate seems pretty solid at about 10 percent before that.
I find the evidence he presents inconclusive at best. If anything, I find it encouraging - no huge sea changes in the ability to start a business in the US over the last 30 years, except for macroeconomic factors.
Yes, we read the same article. What I'm saying is misleading is the chart itself. The chart in the article supports a totally different conclusion than an inflation-adjusted chart. It doesn't actually support "Best economy since the 90s," and yet the author included it. That in itself is misleading to me, whether or not you hedge with a quick statement about inflation afterwards.
1 & 3) seems to be more indicative of the fact that there was a global economic crisis at the end of 2008. I'm not following these points from the paper.
2) seems to point the finger at social media more heavily though I think, but I'm not entirely sure what to make of that but this point is convincing to me.
From the original linked article: "It explains the jobless recoveries of the past and how each recent economic cycle produces higher money figures, yet lower employment. It explains why we are seeing debt driven events that circle the globe."
The graph and data don't explain it, they merely demonstrate it.
Those are not trends, which are comparisons for the same country over multiple years. What the article does and what you're describing are cross-country comparisons in a single year, just at the opposing tails of the distribution.
I love how in an article describing how data can be manipulated to meet any conclusion, Matt cites that the stock market has on average gone up 10% every year and real estate has gained 1% since the great depression.
Who here has money invested from the original great depression? Anyone?
I get his point, and agree with the entire piece but using one set of data to show how using another set of data can be misleading is ... silly.
The central theme of this article does not appear to be well supported by the data presented. I interpret the scatter plot a bit differently they did:
1) The average amount of corporate borrowing as a share of total assets appears remarkably consistent over the last 66 years.
2) The average amount of capital expenditures as a share of total assets appears to be have been declining at a roughly consistent rate for 66 years.
3) The data does not appear to support any firm conclusions regarding changes to the correlation between capital expenditures and borrowing, which was the main topic of the article. The slopes of those regression lines are highly suspect given the amount of data and the outliers in later years. Certainly, there is nothing in the presented data suggesting a change over the last 17 years, much less the last 2 years.
It would have been nice to see some of the raw data or the statistical analysis, but clearly that is for a different audience than what they were targeting.
Stupid article. Takes an exponential trend which it doesn’t show, and points at the fact that the linear version of it over the past 30 years is just a few percentage points per year.
Considering the noise in that chart (savings rate drops from 11% to 5% in one month?!?), I would question whether it's really a trend and how accurate the data is.
One of my pet peeves with economic data - some leading indicator plummets after being high for years. Mass panic in news headlines, only for it to return to normal the next month.
The article may have an interesting premise, however the only statistics we are offered in support of the author’s claim is two years’ worth of data (2021 and 2022). What about the more long-term trends? We might be missing the big picture for short-term cyclical trends. Also, the data for 2022 is, without doubt, incomplete.
I still think having only two data points (one during the Great Recession and one in 2018) reeks of cherry picking. I’d like the author to have provided more historical data to demonstrate the effect.
reply