Home » Analyses
Category Archives: Analyses
Studies show that people do a poor job of guessing how their economic situation compares to other people (example studies here, here, and here). This situation makes sense. It is impolite to pry into other people’s finances. Most people’s knowledge about what constitutes a “high” or “low” income comes from the brackets they read when filing their taxes.
One way to get a more precise sense of one’s own position on the economic ladder is to examine its distribution across society-at-large, which can be done using data from the Survey of Consumer Finances.. Our focus will be on differences in households, as opposed to people, because most people pool their money and expenditures at this level of organization. People’s personal fortunes are often highly dependent on their household’s economic situation. We think of a household’s place on this ladder as a matter of two variables: income (how much money flows into the household regularly) and wealth (the value of the household’s property, less its debts).
In 2019, the median American household earned $58.644 The middle 50% – between the 25th and 75th percentiles – earned between $30.543 and $107.717. An income of at least $191,605 puts you in society’s top 10%, one of $289,960 puts you in the top 5%, and one of $867,620 makes you a One-Percenter in terms of income. To be in the bottom 10%, one’s household had to take in less than $16,290
The bar chart below gives a more detailed sense of the distribution of household incomes.
The 2019 SCF data suggests the median U.S. household net worth of $121,511 The middle 50% (between the 25th and 75th percentiles) are worth between $12,436 and $403,358 More than 10% of American households have negative net worth, which is to say that they owe more money than they own in property. A net worth of $1.2 million puts you in the top 10%, one of $2.6 million puts you in the top 5%, and one of $11 million is enough to be part of the One Percent.
The bar chart below describes the distribution of net worth among U.S. households:
From Broad Categories to Specific Ideas
People have a poor idea of where they stand in relation to society at large. Most wealthy people think that they are middle class. These figures give readers are more specific understanding of where they stand in relation to others in terms of income and wealth. Of course, these benchmarks do not account for age, region, or a host of other factors. If there’s interest, I can break this out further.
Scripts and Data on OSF.
Photo Credit. Glackens, L. M. , Artist. The rich child’s fourth / L.M. Glackens. , 1911. N.Y.: Published by Keppler & Schwarzmann, Puck Building. Photograph. https://www.loc.gov/item/2011649039/.
In a previous post, I examined differences in federal spending across states. This post disaggregates this spending to develop a more detailed view of why different states receive different spending within the current federal compact. I used this data for the analysis. The Markdown file used to generate this analysis is here.
What drives these differences? One way to engage the question is by disaggregating these expenditures. Federal accounting uses five major categories to classify federal spending: direct payments, grants, wages, and procurement. Direct payments are unreciprocated government payments to households, like Social Security, Medicare, welfare payments, or retirement benefits for the military or federal government workers. They occur when the government sends money to a household without getting something in return, like labor, supplies, or some other good or service. Grants are transfers to state/local governments and private enterprises for the fulfillment of some federal government-defined prerogative, including but not limited to Medicare, federal housing aid, roads, public transit, and much else. Procurement expenditures include payments for the federal government’s acquisition of goods and services for both civilian and military purposes. Wages are payments to federal workers, both civilian and military. Both occur where the government sets up operations that require people and supplies.
Direct payments are government payments directly to households. They are comprised mainly of Social Security payments, Medicare coverage (where the federal government covers household medical payments directly), and payments to former federal employees (including military). Per capita direct payments were highest in Vermont ($10,211 per capita), West Virginia ($9,745), Maine ($8,937), Florida ($8,802), Alabama ($8,588), and Mississippi ($8,454). These payments are about one-fifth lower in major donor states, like California ($6,097), New York ($7,068) New Jersey ($7,055), or Massachusetts ($7,191).
Figure 1 describes the distribution of these expenditures (on a per person basis) across states.
Direct payments are primarily comprised of Social Security and Medicare payments (collectively, “federal entitlements), and these state-level differences seem to be primarily driven by differences in how much of these state populations are eligible for these program’s benefits. Federal redistribution between households is primarily focused on helping elderly households. Even wealthy elderly Americans enjoy considerable redistributive transfers. These differences in direct payments are likely to be substantially driven by differences in the prevalence of elderly people in state populations. States like Florida, Arizona, Maine, or Oregon appear to be drawing retirees, who are bringing their government checks with them.
Additionally, reports of state populations using the Social Security program as a form of federally-funded welfare could also be at work here. For example, approximately 9% of West Virginia’s working-age population was reported to be on Workers’ Disability in 2011. Arkansas, Alabama, Kentucky, Mississippi, and Maine all have more than 7% of their working-age population eligible for disability aid, compared to a national average of 4.6% and a considerable number of states with less than 4% disability rates. These disability rates have risen as welfare rolls have been cut.
In addition to these differences in entitlement disbursements, these differences in direct payments are also attributable to differences in benefits for current and former federal employees, including the military. These expenditures are particularly high in the national capital region (Maryland and Virginia), but also across the South and Interior. This is only one part of the unequal distribution of federal jobs that appears in the data, but on its own it amounts to can amount to $1000 or more in per capita federal transfers. For a state like New York, just this small slice of federal transfers related to employment would amount to more than a $19 billion annual infusion into the New York State economy.
Grants are transfers to state and local governments, and non-governmental agencies, to fulfill some federally-defined prerogative. The bulk of this spending is on poor aid. Most of this grant money is for Medicare, and a considerable amount of the remainder goes to TANF, Section 8, the ESEA, Special Education funding, and CHIP. Other grants fund infrastructure development, research, and a range of other purposes. In 2018, the federal government funded 1,274 such grants1.
These are the federal programs that can be of particular benefit to wealthier states, as they fund the programs that underwrite the livelihood of the urban poor, upon whom problems like economic segregation, unaffordable housing, and other problems can weigh more heavily. Such programs also help states with proportionally large poor populations.
Federal operations – whether through the military or in civilian government – can sustain community economies. A federal military base, prison, or other federal enterprise provides the jobs that can help make communities economically viable. To the extent that such jobs are located in communities for the purposes of providing jobs to poor communities, then it is a form of redistribution like any other. It is the government taking money from one community and investing it in another for the purposes of underwriting the economic sustenance of the recipient community.
The figure below shows state differences in combined federal wages (payments to employees for current work) and procurement (payments for supplies, materials, and subcontracted services to sustain federal operations) for both civilian and military operations. We exclude the national capital region states (DC, MD, and VA) because they are far higher than other states and as a result, they distort our scale.
The federal government disproportionately hires and sources in the U.S. South and Interior, and appears to have sidelined states in the Midwest and Mid-Atlantic. States like Alabama and Mississippi receive roughly five times as much spending on jobs and procurement.
These transfers are considerable, amounting to about as much as entitlements. In effect, states who lead the charge against government redistribution primarily live off federal transfers, but those transfers are in the form of jobs that can be construed as “deserving” or proper use of communal resources. However, to the extent that America’s defense requires that military bases be put in Mississippi or Missouri as opposed to upstate New York or Pennsylvania, then these expenditures are being directed towards underwriting some community’s livelihood using money taken from another community.
This exercise helps clarify some finer details explaining how different expenditures contribute to the economic redistribution that occurs through federal government operations. They also help clarify why the states of metropolitan New York experience a large balance of payment losses through the federal government. Our current compact directs federal jobs to the U.S. South and Western Interior, and to states that can attract the elderly, the favored children of American socialism.
Many of the programs that benefit New York – like poor aid or transportation development – are ghettoized in the federal grants system, where they are routinely subject to cutback pressures by the representatives of states who benefit from non-grants redistribution.
This analysis examines the distribution of elderly Americans by state. It is motivated by a broader analysis on federal redistribution by states through fiscal policy, and questions about the degree to which differences in federal direct payments are strongly influenced by the prevalence of older people who are eligible for Social Security and Medicare.
This analysis uses data from the five-year sample of the 2018 American Community Survey. This post offers a quick primer on how to access and analyze the set. Additional tips on analyzing the data can be found in this Markdown file, which you can view here. You can download a table of migrant population data here (or its archive on OSF). If you use the set in your own analysis, please cite this post.
Prevalence of Elderly People
The mean state had an elderly to total population ratio of about 14% elderly, ranging from Alaska (9%) to 18% (Florida). A full table of state scores is appended to this report. These amount to considerable differences. States like Florida, Maine, or West Virginia have double the elderly population of states like Utah or Alaska. Figure 1 (below) presents a map that visualizes these differences in elderly prevalence rates. Refer to the data for detailed estimates.
These differences could conceivably be driven by fertility, mortality, or migration. In other words, either people in states are having more babies (leading to a proportionally young population), living longer (leading to more living older people), or the old and young are disproportionately migrating in or out of states.
Inter-State Relocation by Elderly
We can approximate where the elderly are migrating through the ACS’s question about where respondents lived one year ago. The question can give us a sense of how many people have moved to a state recently. We use the metric as a proxy for measuring states with large elderly populations that migrated for retirement. Figure 2 (below) shows a map of inward elderly migrants in the past year, as measured by the ratio of people aged 65+ who migrated in the past year to the total population.
Note that these are fractions of percentage points relative to the total population. So Arizona and Florida received an estimated 0.5% of their population in inward elderly migration annually, whereas New York only received 0.07%, Illinois 0.8%, California 0.8%, and Texas 0.1%.
At first glance, these might seem like very low numbers. However, keep in mind that these are annual migration rates, and their effects compound. In a state with 10 million people, Arizona’s rate amounts to over 511,000 inward migrants over 10 years, as opposed to over 68,000 at New York’s rate. In fact, Arizona and California are estimated to have had roughly the same absolute number of annual inward migrants, despite the fact that the latter is a much larger state.
Figure 3 (below) describes the size of migration outflows of elderly people, relative to the size of the total population. Remember that this is a ratio of elderly out-migration to the total population, not the elderly population. As such, this metric is interested in general population loss as a result of elderly emigration. It is not a metric of proportionally how frequently seniors leave the state. We will put aside that question for another post.
To get a sense of the magnitude of these differences, consider our hypothetical state with a population of 10 million people. At Vermont’s rate of elderly out-migration (0.42% of the total population leaving annually), that state would lose about 431 thousand people over 10 years. At Texas’s rate, this loss is 87 thousand
My first impression of the data is that smaller Northern states have the highest rate of out-migration. I postulate that small-town, cold communities are more likely to lose elderly residents where the following may post quality-of-life problems: like weather, local healthcare quality, availability of cultural and entertainment opportunities, quality of public transit, walkability, or desirable part-time work opportunities.
After these states, the next largest out-migration (relative to population) are Florida and Arizona. This might be the “buyers’ remorse” out-migrants, or it may be that people who spend their working lives in these states tend to leave the state after their work years. These proportions might be elevated because these states have proportionally larger elderly populations, so their migration would have a more exaggerated impact relative to the total population, which is the denominator we are using here.
Interestingly, states dominated by high-tax, high-cost, and dense urban areas – like New York, New Jersey, Massachusetts, or Illinois – experience median population loss rates. California is near the bottom of elderly out-migration flow size, along with mostly warm weather, low-cost, and low-income states Louisiana, Alabama, and Mississippi, where people may lack the weather incentive or means to emigrate.
Where does this leave us in terms of the balance of migratory flows? The figures below describe net elderly migration, calculated as total inward elderly migration less outward migration, divided by the total population. The result renders a familiar finding – that the Northeast is experiencing a net population loss among the elderly.
Are People Fleeing New York Due to Big City Problems and High Taxes?
The Northeast’s population loss is often described as a matter of people fleeing crumbling cities and high taxes. However, these states are experiencing average levels of out-migration among the elderly. Moreover, New York’s out-migration has been reported to be driven by upstate emigration, not emigration from the cities. This would make sense, given that other parts of the small town North are experiencing similar losses.
From this vantage point, the main issue seems to be low inward migration of the elderly. If someone doesn’t establish a foothold in these communities during their working years, they are unlikely to move here.
Photo Credit. Highsmith, C. M., photographer. (2018) Southwest-style home in Sun City, a historic suburb of Phoenix, Arizona. United States Sun City. Arizona Sun City, 2018. -03-29. [Photograph] Retrieved from the Library of Congress, https://www.loc.gov/item/2018663534/.
This analysis examines the distribution of total federal spending by state.
Distribution of Federal Spending
We begin with a look at the distribution of total expenditures by state. The distribution of expenditures is far more equal than that of federal revenues. Particularly states with particularly high expenditures enjoy roughly double the
Five states receive particularly high levels of federal transfers. Foremost among them are the states of the national capital region – Virginia ($18,678 per person) and Maryland ($18,505) – which receive a disproportionate share of federal spending. Spending is also particularly high in the two non-continguous states: Alaska($18,046) and Hawaii ($15,213). Finally, New Mexico ($16,852) appears to enjoy particularly high spending. The remainder of the states receives between $7,800 (Utah) and $15,012 (Vermont) per person.
Figure 1 (below) presents a map depicting this distribution, minus these extraordinary four states. A full table with precise estimates is appended to this report.
Although per capita expenditures seem more equitable in comparison to the wild variance in federal revenues, these differences in per capita spending amount to a lot of money in the aggregate. A thousand dollar difference in per capita spending is not chump change. Were New Jersey ($10,304 per capita expenditures received) to receive the same spending levels as Missouri ($12,787), more than $22 billion in additional spending would be infused to its economy annually. Such spending amounts to an additional 3.6% of gross state product annually.
My exploration of the data suggests that the most powerful and parsimonious explanation of inter-state differences in federal spending is presented below in Table 1. Diagnostics suggest that, in addition to the capital region states (MD and VA), Utah and Alaska are also outliers. Alaska has very high spending relative to what the model would predict, and Utah is an outlier for low federal expenditures.
This model suggests that, at a baseline, states receive $19,145 per person, assuming zero population and zero poor. Moreover, it suggests that states receive an extra $245 per person for each additional percentage point of the population living below the poverty line. This poverty premium amounts to an additional $1,862 for New Hampshire at 7.6%, the state with the lowest rate. In Mississippi, which leads the nation at 19.8%, the poverty premium to federal spending amounted to an additional $4,851.
The model also predicts a population penalty to government spending that amounts to about $7 per person for each additional percent higher population1. So, for example, in 2018, New York State’ had a population of 19,542,209 was 547% that of Connecticut’s population of 3,572,665. The population penalty levied against New York State is $11,675 per person, compared to $10,493 in Connecticut. New York’s larger population results in an additional spending penalty of $1,182 per person.2. In other words, our analysis shows that federal spending is systematically lower in larger states.
This penalty amounts to considerable stimulus for small states. The predicted difference in federal expenditures between Wyoming (558 thousand people) and California (40 million people) amounts to a 6800% population difference, and a predicted spending difference of $2,937 less spending per person 3. The federal government’s penchant to prefer small states provides these states with a considerable economic boost. Were California to receive Wyoming’s small state funding advantage, California would be receiving on the order of $116 billion in additional spending to the California economy annually.
The Poverty Premium and Population Penalty
This analysis of federal expenditures suggests that the distribution of federal spending across states is influenced by at least two important factors: population size and the prevalence of poverty. This vantage point offers us more insight as to why federal policies shift money away from the union’s donor states. Analysts and commentators routinely point to taxes as a source of these balance of payment deficits, and the comparatively wide variation in tax remittances suggests that it is the biggest factor. However, unequal spending is also a factor. The union’s more donor states – California, New York, New Jersey, Massachusetts, Illinois, and Minnesota – have comparatively low poverty rates and larger populations.
This “population penalty” in spending suggests that the unequal distribution of power in the federal government that favors small states is being leveraged to provide surplus economic transfers to small states. Metropolitan New York has around 21 million people and is represented by six Senators and a few dozen House members. The union’s 16 lowest population states4 also amount to about 21 million people but are represented by 32 Senators and many more Representatives. This structural disadvantage in power over federal government decisions is translated into economic transfers.
Photo Credit. Cary, William De La Montagne, Artist. Emigrants to the West / W.M. Cary. The United States, ca. 1880. [Published 1881] Photograph. https://www.loc.gov/item/92505911/.
This analysis examines how federal taxation and expenditures result in net inflows or outflows to difference U.S. states. The analysis uses our Data on U.S. Federal Balance of Payments, 2018 data set. Figure 1 (below) describes these inter-state money flows resulting from federal policies. A table with presice estimates are also presented at the end of this post..
The data suggests that six (maybe seven) states transferred considerable income to other states in 2018: New Jersey, Massachusetts, New York, Minnesota, Illinois, and California. An addition six states are within $500 in per capita transfers. The majority of U.S. states receive thousands of dollars in per capita transfers from our six “donor states”. These differences are examined through data and analyses detailed in my Federal Redistribution project materials. They describe how a great deal of this redistribution is the result of more tax payments from states that house the metro areas containing the country’s higher income households and businesses. However, there is also inequality in federal spending that is less rooted in the prevalence of poverty or some other need for redistribution, and more in smaller states’ use of political power within our union.
Appendix: Per Capita Balance of Payments by State, 2018
Balance of Payment
Photo Credit. Henle, Fritz, photographer. Las Vegas, Nevada. Transmission towers and transformers redistributing power from Boulder Dam to Basic Magnesium Incorporated, which produces huge quantities of the lightest of all metals for aircraft and other wartime manufacturing. Dec. Photograph. Retrieved from the Library of Congress, <www.loc.gov/item/2017866539/>.
This analysis examines the structure of federal taxes through an examination of state-level federal tax payments data that is disaggregated by tax type. A look at this data provides a more concrete and precise estimate of which communities serve as the economic engines upon which federal government operations depend. This analysis uses a compiled set of state-level public finance and socioeconomic variables detailed here.1 The data is from 2018.
Deconstructing Federal Taxes by Tax Type
Federal taxes are mostly comprised of income taxes. The largest source is personal income taxes, levied on people’s new money receipts. This is a progressively redistributive tax, in the sense that richer people pay proportionally more of their incomes. These yields will be higher in states that house proportionally more very high income households.
The second largest tax source is payroll taxes – the Social Security and Medicare deductions on people’s paychecks. Although these taxes are sometimes considered a form of personal savings or investment in some public retirement fund, the reality of the program is that the services that these taxes fund are financed in the year that they were drawn. It is a tax on current earners to finance benefits by current recipients, and thus the same kind of taxation for current redistribution as any similar redistribution program. These taxes are levied on middle class tax payers, as the rate only applied to the first $132,900 in income, and so their burden is proportionally lighter on those who earn more than this amount, and become proportionally lighter as one earns more money.
The third major income tax is corporate income taxes. This rate is flat. These tax yields will be higher in states where corporations make more money. Combined these income taxes comprise 90% of federal tax receipts.
THe remainder of federal revenues come from a variety of sources. Excise taxes are special taxes on goods like motor fuel, airline tickets, tobacco, and other targeted products. Estate and gift taxes are taxes on larger personal transfers, which in 2018 was in excess of $5.6 million. In 2020, these taxes were on gifts and estates over $11.4 million. Customs are taxes on imports. The remainder is comprised of an assortment of special taxes, fees, and other revenue sources.
Figure 1 (below) presents a pie chart depicting this distribution:
Where are these Taxes Collected?
Below, our analysis focuses on state-level differences in per capita tax payments by these major federal tax types. The analysis gives us a refined sense of which communities generate the revenues that sustain government operations.
Personal Income Taxes
Personal income taxes are progressive,in that wealthier people are supposed to pay higher tax rates and thus give up proportionally more of their income to taxes than poorer people. For example, an individual earning $30,000 per year would have been expected to pay 12% of their income to federal taxes (before adjustments like deductions, credits, etc.). Someone earning $150,000 would pay 24%, and those earning above $501,300 paid 35% of their income in federal taxes.
Of course, wealthy people are known to have the means to reduce their taxes, and we have all seen media stories about super-rich people who pay proportionally less in taxes than poor people. Despite such stories, America’s highest income households appear to be a major source of revenue for the federal government. According to IRS estimates2, America’s top 0.001% paid about 2.1% of all personal income taxes. Its top 0.01% paid about 9% of all personal income tax revenues, the top 0.1% paid 20%, and the top 1% paid 40% of these taxes. Federal revenues from personal income taxes are very heavily comprised of payments from very high income households.
The figure below shows per capita personal income tax payments by state. A table at the end of this post gives precise estimates. The range of per capita payments is considerable, ranging from $2,302 per person in Mississippi to $8,708 in Connecticut. Oregon was the median state, paying $4,225 per person in federal taxes on personal incomes.
Per capita personal income taxes are higher in states whose populations include more high-income earners. Personal taxes are highest in states whose populations are dominated by residents of the country’s high-income and high value-added major metropolitan areas: New York, Boston, Los Angeles, San Francisco, Philadelphia, Chicago, and the District of Columbia regions. Wyomng and South Dakota are outliers among the states that remit large personal taxes on a per capita basis, in that they are not connected to such “mega-cities”. It may be that these states high aggregate payments reflect these states’ very small populations and their success in attracting wealthy residents. Jackson Hole is reputed to be a venue for vacation property ownership among the very wealthy. South Dakota established niche industries in the financial sector. Their small populations could make these cases similar to those of, say, Barbados or the Cayman Islands when examining international differences in macro-finance metrics.
Payroll taxes are mainly comprised of people’s Social Security and Medicare payroll deductions. Unlike personal income taxes, payroll taxes fall most heavily on the upper-middle class, while their impact on very high income households is more limited. In 2018, these taxes were due on the first $128,400 of wages. As such, a person earning $1 million and someone earning $150,000 would pay the same amount in payroll taxes. The rate was the same for all those earning below $128,400
States tend to pay more of these taxes on a per capita basis when their wage-earning population is generally earning more. The figure below demonstrates how per capita payroll tax receipts are higher in states with higher median incomes and lower poverty rates.
It is also worth noting that population concentration in major metro areas and payroll taxes are also higher, because wages are generally higher in highly-productive and -profitable metro regions.
Corporate Income Taxes
Per capita corporate income taxes are the total income taxes levied on a state’s corporation’s income, divided by the total population. The figure gives the reader a sense of how the business activity generated by a state also sustains federal government finances. Tax hauls seen to be higher in states where higher personal incomes prevail. Ultimately, there are a set of states who are experiencing high production or profit, which creates both high income people and enterprises.
The federal government’s finances are primarily sustained by tax levies on the personal incomes of higher income households, and secondarily by taxes on the incomes of wage earners and corporations. All three tend to prevail in states whose populations are concentrated in “rich states” that house the country’s major metropolitan economic powerhouses.
Insofar as taxation levels across states are concerned, US federal government finances looks quite progressive, in that it redistributes money away from communities with high earners and profitable businesses to communities with low wages, less business activity, and less attraction to the wealthy. The benefits of this redistribution to those who live in communities sidelined by the rich and big businesses is obvious – it allows their communities to finance federal services that are better funded than what local taxes could afford on their own. It is progressive in the sense that it takes money from economically dynamic communities to those where such dynamism is in shorter supply.
However, what about the poor who live in these wealthy regions. Are their interests necessarily ensured by this redistribution? It depends on whether the money that is drawn from their community is spent on federal services that benefit the typical, run-of-the-mill family living in metro New York, Los Angeles, Boston, or so on.
|New Jersey||7418||New Hampshire||7418||Massachusetts||7418|
|New Mexico||2876||West Virginia||2876||Montana||2876|
|West Virginia||2577||New Mexico||2577||New Mexico||2577|
- Cohen, Joseph N. 2020. “State Balance of Payment Data, 2018.” Retrieved (<osf.io/eh2d9>).↩
- IRS (2020) “Number of Returns, Shares of AGI and Total Income Tax, AGI Floor on Percentiles in Current and Constant Dollars, and Average Tax Rates, 2001 – 2018” Online data set. https://www.irs.gov/pub/irs-soi/18in01etr.xls↩
This analysis examines the distribution of federal tax payments by state. It uses a compiled set of state-level public finance and socioeconomic variables detailed here. Our data is from 2018. The Markdown file used to generate this analysis can be downloaded from here.
Total Federal Tax Payments
The map below depicts state contributions to total tax revenues, the total amount of money transmitted from people and enterprises in that state to the federal government. See Table A-1 at the end of this post for precise estimates.
States’ total contribution to the U.S. federal government budget is almost perfectly predicted by their population and general income levels. In and of itself, this observation seems unremarkable. However, the map does speak to the federal government’s financial dependency on large state contributions. Half of all federal receipts are generated in eight states – California, Texas, New York, Florida, Illinois, Pennsylvania, New Jersey, and Massachusetts. By contrast, the 25 states who remit smaller aggregate contributions to the federal government generate about 15% of revenues. Although the architecture of the federal government privileges small states in political affairs, federal government financing is highly dependent on the resources provided to it by large states. Large states are important to the Union because they contain most of the people and generate most of the wealth. A policy environment that does damage to large states is one that damages the economic engine of the federal union.
Per Capita Payments
Per capita tax revenues depicts the average tax bill paid by people in a given state. Figure 2 (below) depicts the distribution of per capita federal taxes by state. This metric is calculated by dividing total tax revenues by the state’s population. More precise estimates presented below on Table A-1.
The figure shows that per capita tax payments are higher in the Northeast and West Coast than other states. What explains these per capita tax differentials? The models depicted below in Table 1 suggest that these differences are largely a product of differences in state incomes.
The model predicts that a 10% increase in per capita GDP will result in roughly 8.4% higher per capita tax payments (a very close association, which drives this model’s impressive R-squared). A similar rise in median incomes are predicted to result in a roughly 3% rise in per capita taxes.
What explains these findings? Likely high-income taxes high-income people and businesses, and to a lesser extent by higher payroll taxes yields from areas with higher prevailing incomes.
Overall, this look at the data suggests that federal tax differentials are mainly driven by the distribution of high-income households and businesses. Federal tax receipts are higher in places that house rich people. They are also higher when the typical person enjoys a higher income. Both average and median incomes are related to less poverty as well, but higher living costs.
Table A-1: State Rankings in Total Federal Taxes, Per Capita Taxes, Per Capita GDP, and Median Income, 2018
|State||Total Taxes||State||Per Capita Taxes||State||Per Capita GDP||State||Median Income|
|New York||253,840||New Jersey||13,095||Connecticut||78,312||New Jersey||87,726|
|New Jersey||116,659||California||11,159||North Dakota||74,054||Utah||84,523|
|North Carolina||81,721||South Dakota||10,476||Hawaii||65,541||Illinois||74,399|
|South Carolina||35,891||Wisconsin||9,076||Wisconsin||58,063||Rhode Island||70,151|
|Mississippi||16,850||North Carolina||7,870||Michigan||52,202||South Carolina||62,028|
|New Hampshire||15,800||Utah||7,633||Missouri||51,898||North Carolina||61,159|
|Delaware||10,696||South Carolina||7,059||New Mexico||47,761||Georgia||56,628|
|South Dakota||9,242||Alabama||6,773||South Carolina||46,279||Kentucky||55,662|
|North Dakota||7,418||Idaho||6,660||Idaho||45,086||West Virginia||53,706|
|Alaska||6,872||New Mexico||6,181||West Virginia||42,990||New Mexico||53,113|
- Cohen, Joseph N. 2020. “State Balance of Payment Data, 2018.” Retrieved (<osf.io/eh2d9>).↩
The US Federal Balance of Payments data set is a compilation of state-level public finance, demographic, geographic, and other data, covering the year 2018. This document outlines the variables contained in this set, and the sources from which they were drawn.
The scripts and raw data used to consolidate this data are available for download on this project’s archive on the Open Science Framework.1 The script used to generate these data can be downloaded on Github.
Data & Sources
This section describes the data set and the sources used to build it.
Federal Balance-of-Payment Data
|bop.total||State’s federal balance-of-payments|
|bop.pc||Per capita balance-of-payments|
|tax.paid.total||Taxes paid, total|
|tax.paid.pc||Taxes paid, per capita|
|tax.paid.pinc||Taxes paid: Personal, total|
|tax.paid.pinc.pc||Taxes paid: Personal, per capita|
|tax.paid.socins||Taxes paid: Social Insurance, per capita|
|tax.paid.socins.pc||Taxes paid: Social Insurance, total|
|tax.paid.corp||Taxes paid: Corporate, total|
|tax.paid.corp.pc||Taxes paid, Corporate, per capita|
|tax.paid.excise||Taxes paid: Excise, total|
|tax.paid.excise.pc||Taxes paid, Excise, per capita|
|tax.paid.estate||Taxes paid, Estate, total|
|tax.paid.estate.pc||Taxes paid: Estate, per capita|
|exp.pc||Expenditures, per capita|
|exp.dirpmt.total||Expenditures: Direct payments, total|
|exp.dirpmt.pc||Expenditures: Direct payments, per capita|
|exp.dirpmt.socsec.tot||Expenditures: Direct payments: Social Security, total|
|exp.dirpmt.socsec.pc||Expenditures: Direct payments: Social Security, per capita|
|exp.dirpmt.medicare.tot||Expenditures: Direct payments: Medicare, total|
|exp.dirpmt.medicare.pc||Expenditures: Direct payments: Medicare, per capita|
|exp.dirpmt.vetben.tot||Expenditures: Direct payments: Veterans’ Benefits, total|
|exp.dirpmt.vetben.pc||Expenditures: Direct payments: veterans’ benefits, per capita|
|exp.dirpmt.fedempl.tot||Expenditures: Direct payments: federal employment, total|
|exp.dirpmt.fedempl.pc||Expenditures: Direct payments: federal employment, per capita|
|exp.dirpmt.snap.tot||Expenditures: Direct payments: SNAP, total|
|exp.dirpmt.snap.pc||Expenditures: Direct payments: SNAP, per capita|
|exp.dirpmt.eitc.tot||Expenditures: Direct payments: EITC, total|
|exp.dirpmt.eitc.pc||Expenditures: Direct payments: EITC, per capita|
|exp.dirpmt.ssi.tot||Expenditures: Direct payments: SSI, total|
|exp.dirpmt.ssi.pc||Expenditures: Direct payments: SSIR, per capita|
|exp.grants.total||Expenditures: Grants, total|
|exp.grants.pc||Expenditures: Grants, per capita|
|exp.grants.medicaid.tot||Expenditures: Grants: Medicaid, total|
|exp.grants.medicaid.pc||Expenditures: Grants: Medicaid, per capita|
|exp.grants.hiways.tot||Expenditures: Grants: Highways, total|
|exp.grants.hiways.pc||Expenditures: Grants: Highways, per capita|
|exp.grants.housing.tot||Expenditures: Grants: Housing, total|
|exp.grants.housing.pc||Expenditures: Grants: Housing, per capita|
|exp.grants.chldnut.tot||Expenditures: Grants: Child Nutrition, total|
|exp.grants.chldnut.pc||Expenditures: Grants: Child Nutrition, per capita|
|exp.grants.tanf.tot||Expenditures: Grants: TANF, total|
|exp.grants.tanf.pc||Expenditures: Grants: TANF, per capita|
|exp.grants.esea.tot||Expenditures: Grants: ESEA, total|
|exp.grants.esea.pc||Expenditures: Grants: ESEA, per capita|
|exp.grants.sped.tot||Expenditures: Grants: Special Ed, total|
|exp.grants.sped.pc||Expenditures: Grants, Special Ed, per capita|
|exp.grants.chip.tot||Expenditures: Grants: CHIP, total|
|exp.grants.chip.pc||Expenditures: Grants: CHIP, per capita|
|exp.grants.transit.tot||Expenditures: Grants: Transit, total|
|exp.grants.transit.pc||Expenditures: Grants: Transit, per capita|
|exp.proc.total||Expenditures: Procurement, total|
|exp.proc.pc||Expenditures: Procurement, per capita|
|exp.proc.mil.tot||Expenditures: Procurement: Military, total|
|exp.proc.mil.pc||Expenditures: Procurement: Military, per capita|
|exp.proc.nonmil.tot||Expenditures: Procurement: Non-Military, total|
|exp.proc.nonmil.pc||Expenditures: Procurement: non-military, per capita|
|exp.wages.total||Expenditures: Wages, Total|
|exp.wages.pc||Expenditures: Wages, per capita|
|exp.wages.mil.tot||Expenditures: Wages: military, total|
|exp.wages.mil.pc||Expenditures; Wages: military, per capita|
|exp.wages.civ.tot||Expenditures: Wages, civilian, total|
|exp.wages.civ.pc||Expenditures: Wages, civilian, per capita|
State Population & Density Data
We use a Census distribution for state-level population, population growth, and density data.3 To examine the role that large city populations, we used Census data to score the percentage of a state’s population residing in a Combined Statistical Area with at least 2 million people. These variables are all described below:
|pop.growth||Population Growth (%)|
|pop.density||State Population Density|
|metro.pop.pct||Metro Population (%)|
State GDP Data
State GDP data was drawn from the Bureau of Economic Analysis.4 The set includes data on overall GDP, and government value-added. It also breaks down government value-added to federal civilian, state/local civilian, and military.
|gov.va||Total Government Value-Added|
|gov.va.fedciv||Govt Value;Added: Federal Civilian|
|gov.va.mil||Govt Value-Added: Military|
|gov.va.sl||Govt Value-Added: State & Local|
|gdp.pc||GDP per capita|
R’s base package includes data on state’s geographic size and the Census divisions to which states pertain.
|state.area||State area in mi^2|
The set includes two poverty metrics, drown from the U.S Department of Agriculture’s Economic Research Service5
|pov.u18.pct||Under-18 Poverty Rate|
|costliving||State cost of living|
Median Household Incomes
|med.income||Median Household Income|
Citing this Set
Please cite this set as:
Cohen, Joseph N. 2020. “State Balance of Payment Data, 2018.” Data set. <osf.io/eh2d9>
- Cohen, Joseph N. 2020. “State Balance of Payment Data, 2018.” Retrieved <osf.io/eh2d9>.
- New York State Comptroller’s Office (2020) New York’s Balance of Payments in the Federal Budget, January https://www.osc.state.ny.us/files/reports/budget/pdf/federal-budget-fiscal-year-2018.pdf.
- U.S. Census Bureau “Table 14. State Population – Rank, Percent Change, and Population Density” Data set downloaded November 1, 2020 from https://www2.census.gov/library/publications/2011/compendia/statab/131ed/tables/12s0014.xls>
- United States Bureau of Economic Analysis (2020) “Regional Data: GDP and Personal Income” Online dataset. Accessed November 11, 2020. https://apps.bea.gov/itable/iTable.cfm?ReqID=70&step=1#reqid=70&step=1&isuri=1
- U.S. Department of Agriculture (2020) “Home / Data Products / County-level Data Sets / Poverty” Electronic data set downloaded December 9, 2020 from https://data.ers.usda.gov/reports.aspx?ID=17826
- U.S. Bureau of Economic Analysis (2020) “Regional Data: SARPP Regional Price Parities by State” Electronic data set downloaded December 9, 2020 from https://apps.bea.gov/iTable/iTable.cfm?reqid=70&step=1&isuri=1&acrdn=8#reqid=70&step=1&isuri=1
- U.S. Census Bureau (2020) “Table H-8: Median Household Income by State: 1984 to 2019” Historical Income Tables: Households
Household finance is a big topic. There is much to learn. There will be a lot you don’t know after studying the topic for years. In fact, there’s a lot about household finance that no one knows for sure.
When getting started with a thesis or empirical paper on the topic, it is a smart move to narrow your focus. Focus on one facet of household finance, and build out from there in subsequent projects.
Why Study Household Finance?
- It matches your personal interest and aptitudes. The most important reason to study this topic is that talking about money interests you, and you have a natural aptitude for it.
- Good topic for those seeking work in marketing or finance. Marketing and finance are two major industries in New York City. Develop your knowledge of people’s money as part of your attempt to enter these industries.
Focus on What?
I mean “focus” in three ways. Focus on:
- Playing the role of clarifying our collective discussion by generating and interpreting empirics. What distinguishes our program, and your training with us, is a philosophical dedication to the primacy of probing and verifying factual propositions. We do not strongly emphasize philosophy or theory or politics. In any discussion that you confront in your online research or literature review, remember that your strength is to be the one who comes in with data and helps the group establish which assumed facts are likely true or false.
- A particular facet of households’ finances. Choose some facet of income, expenditures, assets, and debts. See below for more.
- A particular group. I recommend waiting until you see some empirics before making that choice. If, for example, you develop a whole study to explain why college grads spend more on books, and then you find that they don’t spend more on books, then it’s going to result in lost time and possibly a late thesis.
Which Facet of Household Finances?
For your thesis, I recommend that you focus on a particular facet of households’ income, expenditures, savings, assets, debt, or wealth. To learn more about these categories, read the first few chapters of my 2017 book (free to download here).
Incomes are concerned with money flowing into a household. Just about any survey tries to capture some kind of income measure, but a household finance specialist would dive into the weeds of how people earn money. Households receive income as a result of work wages, transfers from family members or friends, government payments, investment returns, and much else.
Expenditures concern where households expend money. Note that expending (spending) money is not the same as investing. For U.S. households, the Consumer Expenditure Survey is the best gold standard fine-grained data set. It parses household spending into a wide range of categories — from food to apparel to shelter to entertainment to healthcare and much else, and drilled down to narrow product categories.
Savings represent the unspent portion of a household’s income. It is income less savings. These savings are channeled into asset categories in discussions of wealth (below). The Survey of Consumer Finances offers a view of how Americans save their money.
Assets are property that can be exchanged for money or used to make money. This includes houses, financial investments, vehicles, retirement savings, real estate investments, intellectual property rights, and other property. Focus on the Survey of Consumer Finances.
Debt is money that is owed to others. Common debt vehicles include credit cards, student loans, medical loans, installment purchase loans, car loans, and mortgages. Focus on Survey of Consumer Finances
Net Worth / Wealth
Net worth is assets less debt. It is the value of a household’s owned property were its debts to be instantaneously paid off through asset liquidation. Focus on Survey of Consumer Finances
Pick a Focus
I recommend focusing on one of the above topics. You decide how narrow or broad. The Consumer Expenditure Survey will allow you to go as narrow as figuring out how much money people spend on toothpaste, or who spends the most on it. You can go as broad as asking how much households typically spend on all their healthcare in a year.
Get some ideas, and then return to your professor and colleagues to discuss next steps.
Studying Household Finance at QC MADASR
A household finance thesis might be a good option for a student who is interested in pursuing professional work in fields that engage marketing, finance, or economics. I do supervise theses on these topics, but require that students meet with me to ensure that this type of work is both practical and fits in their career strategy. To schedule a meeting, contact me at email@example.com.
Do big governments make society poorer or politically authoritarian? We have all heard people argue that growing the government will turn us into the Soviet Union, pressing us all into two-hour lines at the local GUM to buy grey clothes and the family loaf of bread. People are quick to explain how public healthcare is another step towards Chavez’s Venezuela. I struggled with these arguments while I studied neoliberalism. In the end, I concluded that these are extremely weak theories that gain credence by virtue of their endless repetition, and the fact that it is easy to get people to defer to you if you make your explanations long, boring, and complicated.
I find most arguments for this belief to be highly speculative, often very abstract, and routinely plagued by slippery slope reasoning. Most importantly, the proposition that societies experience worse political or economic outcomes runs contrary to what is indicated by straightforward empirical comparisons. Always beware of policy arguments that ask you to doubt straightforward empirics in favor of complicated, highly conjectural theories.
Empirically, wealthier and better-governed countries do not have smaller governments. If any relationship exists, it seems most likely that better-developed societies are more likely to channel proportionally more of their economic activity through governments. Consider the following comparisons of government spending, economic production, and governance quality metrics. These are just two of many empirics that can easily be produced. You can access the data here, and the Markdown file used to generate this report can be downloaded here.
This analysis examines the relationship between government spending (% GDP), the ratio of government spending to the value of all annual national economic production (GDP). Countries who score higher on this metric have governments to spend more money and hire more people, relative to the overall amount of economic activity that is occurring in an economy.
I consider two government spending metrics. Central Government Expenditures (% GDP) express the ratio of national-level government annual spending to GDP. These data are drawn from the World Development Indicators, accessed using the WDI R package. These data cover more of the world’s countries, but will underestimate government expenditures in political systems with major sub-national governments (e.g., as in the U.S. and Canada). Our data on General Government Spending (% GDP) combines central, state and local government spending, relative to national GDP. This data comes from the OECD and was accessed using the OECD R package. It covers a smaller sample of mostly wealthy countries.
Our analysis cover 157 countries with populations of at least one million people. Data is from 2017.
First, consider the relationship between government spending and per capita Gross Domestic Product (GDP). GDP roughly measures the money value of the economic production that takes place in a country. A country with higher GDP per person is engaged in higher value-added production activities, a sign that a society is generally wealthier and commands more purchasing power on global trade markets.
The figure below presents a scatterplot of central government expenditures (% GDP) and per capita GDP across our sample.
The pairwise correlation of these metrics is 0.18, suggesting a weak relationship. The graph can be interpreted as exhibiting no trend. However, the relationship could be stronger when we consider that the U.S., Canada, and Germany have decentralized government systems that underestimate the overall public sector’s size. That observation, coupled with the temptation to treat Timor, Lesotho, and Afghanistan as outliers, could press us to see a positive relationship if we squint. But these types of analytical moves are anti-conservative, and a fair and conservative appraisal is that the relationship between government spending and per captia GDP is weakly positive at best. It certainly isn’t negative though – richer countries clearly do not have smaller governments.
If we use consolidated government expenditures, which include state and local governments, the correlation is roughly the same, and the graph imparts a similar general conclusion.
The data make it clear that poorer countries do not have bigger governnments. Many of the countries with relatively larger governments are wealthier.
Below, we make a similar comparison between countries’ government spending and two metrics of governance quality from the World Bank’s World Governance Indidcators, which can also be accessed through the WDI package. I consider two indices. The first is the Voice and Accountability metric, which captures countries’ general rankings on international expert surveys that assess government system’s responsiveness and accountability to popular will. It is measured on a standardized scale, measured as standard deviations from the mean score (set at zero on this scale).
Again, when we look at general expenditures, the relationship holds. If anything more democratic systems have larger public sectors.
Rule of Law
This principle bears out over more development and governance metrics, but we will try one more comparison to establish the basic point that more government does not mean worse government. This comparison looks at the World Governance Indicators’ Rule of Law index, which captures the degree to which the law is universally applicable and evenly applied. Again, the same relationships hold.
Societies Get Bigger Governments as They Develop Economically
These relationships exhibit an empirical regularity known to economists as “Wagner’s Law” after the 19th-century economist, Adoph Wagner, to whom the observation that governments and economies grow together is attributed. Wealthier societies have bigger governments because these economies are more complex, with more moving parts to govern. They are also richer and have the resources to ensure higher living standards for everyone in society. If we are thinking about this issue with a strong empiricism and a leaning towards straightforward explanations, your inclination should be to be credulous about strong theories that government harms freedom and prosperity, unless you want to dwell on extreme and unlikely scenarios.