Bad data, bad decisions... 9 years later after Obamacare, Centers for Medicare and Medicaid Services' actuaries find that CBO numbers were wildly off

9 Years After Obamacare Passed, Agency Finds Numbers Were Wildly Off. Jarrett Stepman. The Daily Signal, Feb 22, 2019. https://www.dailysignal.com/2019/02/22/9-years-after-obamacare-passed-agency-finds-numbers-were-wildly-off

Democrats defeated Republicans in the Obamacare repeal fight by warning that 22 million Americans would be thrown off their health insurance. They pointed to data leaked from the Congressional Budget Office.

Well, it turns out that data was completely wrong.

According to a report [a paywalled paper, https://www.healthaffairs.org/doi/full/10.1377/hlthaff.2018.05499] by the Centers for Medicare and Medicaid Services released Wednesday, the Congressional Budget Office wildly overestimated the number of people who would lose their health insurance with the repeal of the individual mandate penalty.

Initial estimates from the Congressional Budget Office said 14 million would drop off their health insurance coverage due to the elimination of the individual mandate. Then, during the height of the 2017 debate over repeal, progressives touted a leaked number from the Congressional Budget Office claiming that 22 million people would “lose” their insurance if Congress repealed the law.

However, as health care analyst Avik Roy pointed out (https://www.forbes.com/sites/theapothecary/2017/07/22/cbo-three-fourths-of-coverage-difference-between-obamacare-gop-bills-driven-by-individual-mandate/#320be2c63627
), what made this number so high was the inflated number of people expected to lose their insurance due to repeal of the mandate—about 73 percent to be exact. So, it wouldn’t be 22 million Americans losing their insurance. Most of those in the projection would simply be choosing to opt out of insurance.

And it turns out even that wasn’t true. A far smaller number of Americans appear to be opting out of insurance since the individual mandate’s repeal. Only 2.5 million more people are expected to go without insurance in 2019 due to its repeal, according to the latest report, and that number is expected to decline in the years ahead.

[...]

The Congressional Budget Office is opaque, to say the least. It does not publish or share the way it comes up with numbers, and some have criticized the organization for its lack of transparency and outsized influence on policymaking.

Doug Badger, a visiting fellow in domestic policy studies at The Heritage Foundation, told The Daily Signal that Congressional Budget Office analysis has been a chronic problem.

“When it comes to the individual mandate, CBO has never let the facts affect their wildly inaccurate estimates. CBO continued to forecast that millions of insured Americans would suddenly become uninsured if the mandate were repealed,” Badger wrote in an email to The Daily Signal. “CBO’s faulty estimates misled the public into believing that repealing Obamacare would lead to a vast increase in the number of uninsured. Bad estimates produced bad policy.”

[...]


---
the paywalled paper:

National Health Expenditure Projections, 2018–27: Economic And Demographic Trends Drive Spending And Enrollment Growth. By Andrea M. Sisko, Sean P. Keehan, John A. Poisal, Gigi A. Cuckler, Sheila D. Smith, Andrew J. Madison, Kathryn E. Rennie, and James C. Hardesty


doi: 10.1377/hlthaff.2018.05499

ABSTRACT National health expenditures are projected to grow at an average annual rate of 5.5 percent for 2018–27 and represent 19.4 percent of gross domestic product in 2027. Following a ten-year period largely influenced by the Great Recession and major health reform, national health spending growth during 2018–27 is expected to be driven primarily by long-observed demographic and economic factors fundamental to the health sector. Prices for health care goods and services are projected to grow 2.5 percent per year, on average, for 2018–27—faster than the average price growth experienced over the last decade—and to account for nearly half of projected personal health care spending growth. Among the major payers, average annual spending growth in Medicare (7.4 percent) is expected to exceed that in Medicaid (5.5 percent) and private health insurance (4.8 percent) over the projection period, mostly as a result of comparatively higher projected enrollment growth. The insured share of the population is expected to remain stable at around 90 percent throughout the period, as net gains in health coverage from all sources are projected to keep pace with population growth.

-
During 2018–27 national health spendingisexpectedtobedriven primarily by long-observed demographic and economic factors fundamentaltothehealthsector, largely in contrast to the prior decade—which was affected by the notable impacts of a historic recessionand the implementation of wide-ranging health reform legislation.1 Overall, national health spending is projected to grow at 5.5 percentperyear,onaverage,for2018–27(exhibit1). This is faster than the average growth rate experiencedfollowingthelastrecession(3.9percent for 2008–13) and the more recent period inclusive of the Affordable Care Act’s major coverage expansions(5.3percentfor2014–16).However, it is slower than the rate throughout the nearly two decades preceding the Great Recession (7.3 percent for 1990–2007). Growth in gross domesticproduct(GDP)duringtheten-yearprojectionperiodisprojectedtoaverage4.7percent. Because national health spending growth is expectedtoincrease0.8percentagepointfaster, onaverage,thangrowthinGDPovertheprojection period, the health share of GDP is expected to rise from 17.9 percent in 2017 to 19.4 percent in 2027, with almost all of the increase in share expected after 2020. Projected average annual spending growth rates for the underlying major payers of health care are expected to vary substantially during 2018–27,mainlyasaresultofdifferingexpected trends in enrollment growth. Average Medicare spendinggrowthisprojectedtobethefastest,at
doi: 10.1377/hlthaff.2018.05499 HEALTH AFFAIRS 38, NO. 3 (2019): – ©2019 Project HOPE— The People-to-People Health Foundation, Inc.
Andrea M. Sisko (Andrea .Sisko@cms.hhs.gov) is an economist in the Office of the Actuary, Centers for Medicare and Medicaid Services (CMS), in Baltimore, Maryland.
Sean P. Keehan is an economist in the CMS Office of the Actuary.
John A. Poisal is a deputy director of the National Health Statistics Group, CMS Office of the Actuary.
Gigi A. Cuckler is an economist in the CMS Office of the Actuary.
Sheila D. Smith is an economist in the CMS Office of the Actuary.
Andrew J. Madison is an actuary in the CMS Office of the Actuary.
Kathryn E. Rennie is an actuary in the CMS Office of the Actuary.
James C. Hardesty is an actuary in the CMS Office of the Actuary.
March 2019 38:3 Health Affairs 1
Costs & Spending
Downloaded from HealthAffairs.org on February 20, 2019. Copyright Project HOPE—The People-to-People Health Foundation, Inc. For personal use only. All rights reserved. Reuse permissions at HealthAffairs.org.
Exhibit 1
National health expenditures(NHE),aggregateand percapitaamounts,shareofgross domesticproduct(GDP), andaverageannualgrowthfromprevious year shown, by source of funds, selected calendar years 2013–27
Source of funds 2013a 2016 2017 2018b 2019b 2027b Expenditure, billions NHE $2,881.8 $3,361.1 $3,492.1 $3,646.9 $3,823.1 $5,963.2 Health consumption expenditures 2,728.6 3,202.9 3,324.5 3,470.3 3,637.6 5,679.9 Out of pocket 325.9 356.1 365.5 378.6 396.9 585.8 Health insurance 2,088.1 2,504.5 2,604.2 2,720.9 2,850.6 4,545.8 Private health insurance 947.1 1,136.4 1,183.9 1,237.7 1,278.2 1,896.7 Medicare 589.9 677.1 705.9 747.4 800.1 1,436.8 Medicaid 445.2 565.6 581.9 594.8 623.4 992.1 Federal 256.9 358.3 361.2 369.5 386.5 611.1 State and local 188.4 207.3 220.6 225.3 237.0 380.9 Other health insurance programsc 105.9 125.3 132.6 141.0 148.8 220.2 Other third-party payers and programs and public health activity 314.7 342.4 354.8 370.8 390.0 548.4 Investment 153.2 158.2 167.6 176.5 185.5 283.3 Population (millions) 315.7 322.9 325.2 327.9 330.7 352.7 GDP, billions $16,784.9 $18,707.2 $19,485.4 $20,498.6 $21,503.1 $30,755.4 Disposable personal income, billions 12,505.3 14,170.9 14,796.3 15,563.2 16,297.3 23,453.9 NHE per capita 9,128.9 10,410.1 10,739.1 11,121.2 11,559.3 16,907.0 GDP per capita 53,170.5 57,941.2 59,922.8 62,511.0 65,015.9 87,198.3 Prices (2012=100.0) Personal Health Care Price Index 1.015 1.049 1.062 1.081 1.101 1.359 GDP Implicit Price Deflator, chain weighted 1.018 1.059 1.079 1.104 1.130 1.344 NHE as percent of GDP 17.2% 18.0% 17.9% 17.8% 17.8% 19.4% Annual growth NHE 3.9% 5.3% 3.9% 4.4% 4.8% 5.7% Health consumption expenditures 4.0 5.5 3.8 4.4 4.8 5.7 Out of pocket 2.0 3.0 2.6 3.6 4.8 5.0 Health insurance 4.4 6.2 4.0 4.5 4.8 6.0 Private health insurance 3.4 6.3 4.2 4.5 3.3 5.1 Medicare 5.3 4.7 4.2 5.9 7.1 7.6 Medicaid 5.3 8.3 2.9 2.2 4.8 6.0 Federal 5.6 11.7 0.8 2.3 4.6 5.9 State and local 5.0 3.2 6.4 2.1 5.2 6.1 Other health insurance programsc 6.0 5.8 5.8 6.4 5.5 5.0 Other third-party payers and programs and public health activity 3.4 2.9 3.6 4.5 5.2 4.4 Investment 1.7 1.1 6.0 5.3 5.1 5.4 Populationd 0.8 0.8 0.7 0.8 0.9 0.8 GDP 2.5 3.7 4.2 5.2 4.9 4.6 Disposable personal income 2.9 4.3 4.4 5.2 4.7 4.7 NHE per capita 3.0 4.5 3.2 3.6 3.9 4.9 GDP per capita 1.7 2.9 3.4 4.3 4.0 3.7 Prices (2012=100.0) Personal Health Care Price Index 2.2 1.1 1.3 1.7 1.9 2.7 GDP Implicit Price Deflator, chain weighted 1.6 1.3 1.9 2.3 2.3 2.2
SOURCES Centers for Medicare and Medicaid Services, Office of the Actuary, National Health Statistics Group; and Department of Commerce, Bureau of Economic Analysis and Bureau of the Census. NOTES For definitions, sources, and methods for NHE categories, see CMS.gov. National Health Expenditure Accounts: methodology paper, 2017: definitions, sources, and methods [Internet]. Baltimore (MD): Centers for Medicare and Medicaid Services; [cited 2019 Jan 25]. Available from: https://www.cms.gov/Research-Statistics-Data-and-Systems/Statistics-Trends-and-Reports/NationalHealthExpendData/Downloads/dsm-17.pdf. Numbers might not add to totals because of rounding. Percent changes are calculated from unrounded data. Tables with data for all years of the projection period can be found at CMS.gov. NHE projections 2018–27—tables [Internet]. Baltimore (MD): Centers for Medicare and Medicaid Services; 2019 [cited 2019 Feb 20]. Available from: https://www.cms.gov/Research-Statistics-Data-and-Systems/Statistics-Trends-and-Reports/NationalHealthExpendData/Downloads/Proj2018Tables.zip. aAnnual growth, 2008–13. bProjected. cIncludes health-related spending for Children’s Health Insurance Program (CHIP), Titles XIX and XXI; Department of Defense; and Department of Veterans Affairs. dEstimates reflect the Bureau of the Census’s definition of resident-based population (which includes all people who usually reside in the fifty states or the District of Columbia but excludes residents living in Puerto Rico and areas under US sovereignty, and US Armed Forces overseas and US citizens whose usual place of residence is outside of the United States). Estimates also include a small (typically less than 0.2 percent of population) adjustment to reflect census undercounts. Projected estimates reflect the area population growth assumptions found in the 2018 Medicare Trustees Report (see note 4 in text).
Costs & Spending
2 Health Affairs March 2019 38:3
Downloaded from HealthAffairs.org on February 20, 2019. Copyright Project HOPE—The People-to-People Health Foundation, Inc. For personal use only. All rights reserved. Reuse permissions at HealthAffairs.org.
7.4 percent per year, as the shift of the babyboom generation into the program continues toresultinrobustgrowthinenrollment(2.5percent per year, on average) (exhibit 2). This shift also contributes to comparatively slower projected private health insurance enrollment growth of just 0.2 percent per year in 2018–27 and underlies the expectation that growth in private health insurance spending will be the slowest among the payers, at just 4.8 percent peryear,onaverage.Medicaidspendinggrowth is expected to be 5.5 percent, on average, with projected enrollment growth of 1.3 percent per year during this period. Per enrollee, rates of growth in spending for Medicare, Medicaid, and private health insurance are expected to be somewhat similar over the ten-year projection period (4.7 percent,
4.1percent,and4.6percentperenrollee,respectively). However, these averages mask the unique year-to-year trends among the major payersthatareinfluencedbyregulation,legislation, and economic factors—each of which is discussed in more detail below. For 2018, national health spending is projected to have grown by 4.4 percent, following a rate of 3.9 percent in 2017 (exhibit 1).1 Faster projected spending growth of almost 2 percentage points in Medicare (5.9 percent) primarily contributes to the acceleration that reflectshigherexpectedgrowthforbothhospital services and prescription drugs. However, Medicaid spending growth is projected to have slowed by 0.7 percentage point in 2018 (to 2.2 percent), as enrollment growth is expected to have slowed for the fourth consecutive year.
Exhibit 2
National health expenditures (NHE) and health insurance enrollment, aggregate and per enrollee amounts, and average annual growth from previous year shown, by source of funds, selected calendar years 2013–27
Source of funds 2013a 2016 2017 2018b 2019b 2027b Expenditure, billions Private health insurance $947.1 $1,136.4 $1,183.9 $1,237.7 $1,278.2 $1,896.7 Medicare 589.9 677.1 705.9 747.4 800.1 1,436.8 Medicaid 445.2 565.6 581.9 594.8 623.4 992.1 Annual growth in expenditure Private health insurance 3.4% 6.3% 4.2% 4.5% 3.3% 5.1% Medicare 5.3 4.7 4.2 5.9 7.1 7.6 Medicaid 5.3 8.3 2.9 2.2 4.8 6.0 Per enrollee spending Private health insurance $ 5,052 $ 5,771 $ 6,001 $ 6,269 $ 6,511 $ 9,384 Medicare 11,503 12,144 12,347 12,726 13,240 19,546 Medicaid 7,553 7,944 8,013 8,099 8,289 12,029 Annual growth in per enrollee spending Private health insurance 4.3% 4.5% 4.0% 4.5% 3.9% 4.7% Medicare 2.4 1.8 1.7 3.1 4.0 5.0 Medicaid 0.9 1.7 0.9 1.1 2.4 4.8 Enrollment, millions Private health insurance 187.5 196.9 197.3 197.4 196.3 202.1 Medicare 51.3 55.8 57.2 58.7 60.4 73.5 Medicaid 58.9 71.2 72.6 73.4 75.2 82.5 Uninsured 44.1 28.7 29.7 29.9 31.2 36.2 Population 315.7 322.9 325.2 327.9 330.7 352.7 Insured share of total population 86.0% 91.1% 90.9% 90.9% 90.6% 89.7% Annual growth in enrollment Private health insurance −0.9% 1.7% 0.2% 0.1% −0.6% 0.4% Medicare 2.9 2.8 2.5 2.7 2.9 2.5 Medicaid 4.4 6.5 2.0 1.1 2.4 1.2 Uninsured 1.2 −13.4 3.7 0.7 4.3 1.9 Population 0.8 0.8 0.7 0.8 0.9 0.8
SOURCE Centers for Medicare and Medicaid Services, Office of the Actuary, National Health Statistics Group. NOTES For definitions, sources, and methods for NHE categories, see CMS.gov. National Health Expenditure Accounts: methodology paper, 2017 (see exhibit 1 notes). Numbers might not add to totals because of rounding. Percent changes are calculated from unrounded data. Tables with data for all years of the projection period can be found at CMS.gov. NHE projections 2018–27—tables (see exhibit 1 notes). aAnnual growth, 2008–13. bProjected.
March 2019 38:3 Health Affairs 3
Downloaded from HealthAffairs.org on February 20, 2019. Copyright Project HOPE—The People-to-People Health Foundation, Inc. For personal use only. All rights reserved. Reuse permissions at HealthAffairs.org.
Fromtheperspectiveofoverallhealthinsurance enrollment,netgainsinhealthinsurancecoverage across all sources are expected to have kept pacewithoverallpopulationgrowth.Asaresult, the insured share of the population is projected to have remained stable at 90.9 percent. For 2019, growth in national health spending is expected to increase again to 4.8 percent (exhibit 1). Medicare spending growth is projected to continue accelerating (to 7.1 percent), partlyas aresultof fastergrowth inperenrollee spending attributable to higher fee-for-service payment updates. Growth in Medicaid expenditures is also expected to rise (to 4.8 percent), in partbecauseofexpansionsofMedicaidcoverage in Idaho, Maine, Nebraska, Utah, and Virginia. A somewhat mitigating influence on overall national health spending growth, however, is the expected impact of the repeal of the individual mandate. The repeal is expected to result in lowerprivatehealthinsuranceenrollment,since some people—particularly those with directpurchase insurance—may elect to forgo coverage.2,3 Combined,theseshiftsinenrollmentlead to a projected net increase in the number of uninsured of 1.3 million people, to 31.2 million in2019(exhibit 2). However,projectedgains in enrollment through other sources are expected to partially offset those declines, resulting in only a slight decrease in the insured share of the population (to 90.6 percent in 2019, from 90.9 percent in 2018). For2020–27,growthinnationalhealthspendingisexpectedtoaverage5.7percent.Thisrateis fasterthanprojectedfor2019,andfastergrowth is generally evident for the underlying major payers and health care services and goods (exhibits 1 and 3). The acceleration is in part duetofastergrowthinpersonalhealthcarepricesasmeasuredbythePersonalHealthCarePrice Index (exhibit 1). Also contributing is increasingly higher expected growth in utilization on thepartofMedicarebeneficiariesandthosewith privatehealthinsurance,thelatterinfluencedby a lagged response to comparatively higher income growth during 2020–22. With respect to insurance coverage over 2020–27, growth in employer-sponsored health insurance enrollmentis projected to bebelow thatof population growth and decline for those purchasing insurance directly, which contributes to a slight decline in the insured share of the population to 89.7 percent by 2027 (exhibit 2). The share of health care spending sponsored (orfinanced)byfederal,state,andlocalgovernments is expected to increase by 2 percentage points during 2018–27, reaching 47 percent by 2027 (exhibit 4). The increase is entirely accounted for by the federal government share,
which is expected to grow from 28 percent in 2017 to 31 percent in 2027, and largely reflects fastergrowth inMedicare spending as the babyboomgenerationcontinuestotransitionintothe program. The expected business and household share is expected to fall from 55 percent in 2017 to 53 percent in 2027.
Model And Assumptions The national health expenditure projections incorporate a combination of actuarial and econometricmodelingmethods,aswellasjudgments about future events and trends that are expectedtoinfluencehealthspending.3 Theyare largely based on economic and demographic assumptions in the 2018 Medicare Trustees Report,4updatedtoreflectmorerecentlyreleased macroeconomic data.3 The projections also reflect current law5 and do not reflect any policy proposals currently under consideration. Estimates of future health care spending and enrollment are inherently subject to substantial uncertainty that increases over the projection horizon. In addition to the potential effects of evolvinghealthcaremarketsandchangesinlaw over time, economic conditions can differ from the intended midrange assumptions used here. In the case of one economic variable, disposable personal income, analysis by the Office of theActuaryhasconsistentlyfoundarelationship between growth in that metric and growth in health spending, especially for private health insurance.3 That is, as income growth increases or decreases, health spending growth tends to follow in the same direction, but with a lag. Thisrelationshiphasbeenevidentoverthefull history of the National Health Expenditure Accounts and is reflected in these projections.3 As a result, with faster growth in income assumed for the coming decade relative to the recent past, it is expected that health spending growth will respond and be higher as well.3 The projections presented here reflect this relationship. Thus, to the extent that actual growth in income differs from what is assumed, actual growthinhealthspendingmaydifferfromwhat is projected.
Factors Accounting For Growth Inexhibit5averageannualpersonalhealthcare spending6growthisdecomposedtodemonstrate the relative contributions of underlying price growth (economywide and relative personal health care price inflation), use and intensity, population growth, and age-sex mix. During 2018–27 personal health care spending growth is expected to average 5.5 percent, with growth
Costs & Spending
4 Health Affairs March 2019 38:3
Downloaded from HealthAffairs.org on February 20, 2019. Copyright Project HOPE—The People-to-People Health Foundation, Inc. For personal use only. All rights reserved. Reuse permissions at HealthAffairs.org.
in personal health care prices expected to account for nearly half of that growth, on average. Growth in use and intensity is expected to account for just under one-third of the average annual personal health care spending growth, with population growth and the changing age-sex mix of the population accounting for
the remainder. Over specific years within the projection period, however, there are notable trends in prices and the volume and intensity of services, some of which are anticipated to contrast with recent experience. Inflationforhealthcaregoodsandservices,as measured by the Personal Health Care Price
Exhibit 3
National health expenditures (NHE) amounts and annual growth from previous year shown, by spending category, selected calendar years 2013–27
Spending category 2013a 2016 2017 2018b 2019b 2027b Expenditure, billions NHE $2,881.8 $3,361.1 $3,492.1 $3,646.9 $3,823.1 $5,963.2 Health consumption expenditures 2,728.6 3,202.9 3,324.5 3,470.3 3,637.6 5,679.9 Personal health care 2,438.0 2,851.9 2,961.0 3,085.3 3,242.5 5,058.4 Hospital care 937.6 1,092.8 1,142.6 1,193.4 1,254.7 1,961.6 Professional services 759.4 884.0 920.0 962.8 1,013.6 1,541.2 Physician and clinical services 569.6 666.5 694.3 728.0 767.6 1,172.0 Other professional services 78.7 92.4 96.6 100.8 106.1 165.3 Dental services 111.1 125.1 129.1 134.0 139.9 203.9 Other health, residential, and personal care 144.3 173.4 183.1 188.4 196.9 318.6 Home health care 81.4 93.1 97.0 101.8 108.8 186.8 Nursing care facilities and continuing care retirement communities 149.0 163.0 166.3 170.8 178.0 270.7 Retail outlet sales of medical products 366.3 445.6 451.9 468.1 490.5 779.4 Prescription drugs 265.2 332.0 333.4 344.5 360.3 576.7 Durable medical equipment 45.1 51.0 54.4 57.4 60.9 97.8 Other nondurable medical products 56.0 62.7 64.1 66.2 69.3 105.0 Government administration 37.4 44.7 45.0 46.7 49.4 81.0 Net cost of health insurance 174.2 220.7 229.5 247.2 252.0 417.3 Government public health activities 79.1 85.6 88.9 91.1 93.6 123.2 Investment 153.2 158.2 167.6 176.5 185.5 283.3 Noncommercial research 46.7 47.6 50.7 53.5 56.2 83.3 Structures and equipment 106.5 110.6 116.9 123.1 129.3 200.0 Annual growth NHE 3.9% 5.3% 3.9% 4.4% 4.8% 5.7% Health consumption expenditures 4.0 5.5 3.8 4.4 4.8 5.7 Personal health care 4.1 5.4 3.8 4.2 5.1 5.7 Hospital care 5.2 5.2 4.6 4.4 5.1 5.7 Professional services 3.6 5.2 4.1 4.7 5.3 5.4 Physician and clinical services 3.7 5.4 4.2 4.9 5.4 5.4 Other professional services 4.6 5.5 4.6 4.3 5.3 5.7 Dental services 2.2 4.0 3.2 3.8 4.4 4.8 Other health, residential, and personal care 4.9 6.3 5.6 2.9 4.5 6.2 Home health care 6.0 4.6 4.3 4.9 6.8 7.0 Nursing care facilities and continuing care retirement communities 3.0 3.0 2.0 2.7 4.2 5.4 Retail outlet sales of medical products 2.2 6.8 1.4 3.6 4.8 6.0 Prescription drugs 2.0 7.8 0.4 3.3 4.6 6.1 Durable medical equipment 3.3 4.2 6.8 5.5 6.1 6.1 Other nondurable medical products 2.7 3.8 2.2 3.3 4.7 5.3 Government administration 4.2 6.1 0.5 3.9 5.7 6.4 Net cost of health insurance 3.3 8.2 4.0 7.7 2.0 6.5 Government public health activities 3.1 2.7 3.9 2.4 2.8 3.5 Investment 1.7 1.1 6.0 5.3 5.1 5.4 Noncommercial research 1.5 0.7 6.5 5.4 5.1 5.0 Structures and equipment 1.8 1.3 5.7 5.3 5.1 5.6
SOURCE Centers for Medicare and Medicaid Services, Office of the Actuary, National Health Statistics Group. NOTES For definitions, sources, and methods for NHE categories, see CMS.gov. National Health Expenditure Accounts: methodology paper, 2017 (see exhibit 1 notes). Numbers might not add to totals because of rounding. Percent changes are calculated from unrounded data. Tables with data for all years of the projection period can be found at CMS.gov. NHE projections 2018–27—tables (see exhibit 1 notes). aAnnual growth, 2008–13. bProjected.
March 2019 38:3 Health Affairs 5
Downloaded from HealthAffairs.org on February 20, 2019. Copyright Project HOPE—The People-to-People Health Foundation, Inc. For personal use only. All rights reserved. Reuse permissions at HealthAffairs.org.
Index and inclusive of both economywide and relative personal health care price inflation, is projected to play a larger role in the coming decade (averaging growth of 2.5 percent per year for 2018–27, compared to 1.1 percent for 2014–17)andaccountfornearlyhalfofpersonal health care spending growth. This expectation reflects accelerating growth in both economywide inflation and relative personal health care price inflation (or the difference between price growth for personal health care goods and servicesand economywideinflation).Theexpected acceleration in growth in economywide prices occurredprimarily in 2018. From 2019 forward, asteadyincreaseinrelativepersonalhealthcare priceinflationisprojected,ascertainfactorsthat
contributedtolowornegativegrowthinrelative personal health care price inflation since 2011 are anticipated to be less influential in restraining prices over the next decade. Such factors includerisingsensitivitytopricesbyconsumers and insurers, especially for services subject to cost sharing;7 selective contracting by insurers; and improvements in productivity through the useoflower-costprovidersinphysicianoffices.8 Similarly, input price growth, including healthsector wages, is expected to accelerate as downward pressure on provider prices lessens. The average growth rate for use and intensity of services is projected to be 1.7 percent over 2018–27 and to account for about 30 percent ofpersonalhealthcarespendinggrowth(exhib
Exhibit 4
National healthexpenditures(NHE) amounts,average annualgrowthfromprevious yearshown, and percentdistribution, by type of sponsor, selected calendar years 2013–27
Type of sponsor 2013a 2016 2017 2018b 2019b 2027b Expenditure, billions NHE $2,881.8 $3,361.1 $3,492.1 $3,646.9 $3,823.1 $5,963.2 Businesses, household, and other private revenues 1,620.6 1,836.7 1,914.1 2,002.9 2,095.2 3,136.4 Private businesses 580.4 669.1 696.5 730.9 765.1 1,123.2 Household 833.0 942.8 978.6 1,019.9 1,064.1 1,619.3 Other private revenues 207.2 224.7 239.0 252.0 266.0 393.9 Governments 1,261.2 1,524.4 1,577.9 1,644.0 1,727.9 2,826.8 Federal government 752.7 952.4 982.4 1,032.7 1,089.7 1,833.8 State and local governments 508.5 572.0 595.5 611.2 638.2 993.0 Annual growth NHE 3.9% 5.3% 3.9% 4.4% 4.8% 5.7% Businesses, household, and other private revenues 2.8 4.3 4.2 4.6 4.6 5.2 Private businesses 2.3 4.9 4.1 4.9 4.7 4.9 Household 3.1 4.2 3.8 4.2 4.3 5.4 Other private revenues 3.3 2.7 6.4 5.4 5.6 5.0 Governments 5.3 6.5 3.5 4.2 5.1 6.3 Federal government 6.1 8.2 3.2 5.1 5.5 6.7 State and local governments 4.2 4.0 4.1 2.6 4.4 5.7 Distribution NHE 100% 100% 100% 100% 100% 100% Businesses, household, and other private revenues 56 55 55 55 55 53 Private businesses 20 20 20 20 20 19 Household 29 28 28 28 28 27 Other private revenues 7 7 7 7 7 7 Governments 44 45 45 45 45 47 Federal government 26 28 28 28 29 31 State and local governments 18 17 17 17 17 17
SOURCE Centers for Medicare and Medicaid Services, Office of the Actuary, National Health Statistics Group. NOTES For definitions, sources, and methods for NHE categories, see CMS.gov. National Health Expenditure Accounts: methodology paper, 2017 (see exhibit 1 notes). Numbers might not add to totals because of rounding. Percent changes are calculated from unrounded data. Tables with data for all years of the projection period can be found at CMS.gov. NHE projections 2018–27—tables (see exhibit 1 notes). aAnnual growth, 2008–13. bProjected.
Costs & Spending
6 Health Affairs March 2019 38:3
Downloaded from HealthAffairs.org on February 20, 2019. Copyright Project HOPE—The People-to-People Health Foundation, Inc. For personal use only. All rights reserved. Reuse permissions at HealthAffairs.org.
it5).Thisresultcontrastswiththerateobserved during the years immediately following the implementation of the coverage expansions under theAffordableCareAct(2014–16),whenuseand intensity was the dominant driver of personal health care spending growth—representing 2.9 percentage points, or just over half, of the average spending growth rate of 5.4 percent. Initially, these increases were largely influenced by expanding enrollment, followed by faster per enrollee spending growth that likely reflected care provided to the newly insured. Unlike that unique time period, during 2018–27 growth in the use and intensity of medical care isprimarilyinfluencedbytheanticipatedeffects of macroeconomic growth consistent with the longer-run historical relationship.
Outlook For Spending And Enrollment By Payer Medicare Medicare spending growth is projected to have increased 5.9 percent in 2018, compared to 4.2 percent in 2017 (exhibit 1), mainly because of faster per enrollee spending growth(3.1percentin2018versus1.7percentin 2017) (exhibit 2). Increases in Medicare private health plan payments, as well as spending for fee-for-service hospital care and prescription
drugs, underlie the projected acceleration. In 2019 Medicare spending is projected to increase by 7.1 percent, a 1.2-percentage-point acceleration over growth in 2018. Increases in fee-for-servicepaymentratescomparedto2018, along with slightly faster growth in the use and intensity of physician and clinical services, contribute to faster expected growth in perenrollee spending, which is projected to rise to 4.0 percent. Additionally, projected Medicare enrollment growth reaches its peak at 2.9 percent in 2019, up from 2.7 percent in 2018. Over 2020–27 Medicare spending growth is expected to remain highest among the payers, averaging 7.6 percent. Compared to the 7.1 percent increase projected for 2019, this faster average growth is primarily driven by an expectation of a continued rebound in growth in the useandintensityofservicesusedthroughoutthe period that is more consistent with the program’s long-term experience, compared to that of the past decade. By the end of the projection period (2026–27) the expected growth rate decelerates to around 7.0 percent, down from a projection-period peak of 8.1 percent in 2022, as slower increases in input prices— including for hospitals—and anticipated faster multifactor productivity growth lead to smaller payment updates for many Part A services. En
Exhibit 5
Factors accounting for growth in personal health care (PHC) expenditures, selected calendar years 1990–2027
SOURCES Centers for Medicare and Medicaid Services, Office of the Actuary, National Health Statistics Group; and Department of Commerce, Bureau of Economic Analysis and Bureau of the Census. NOTES “Relative PHC inflation” represents the share of medical price growth that exceeds economywide inflation. “Economywide inflation” reflects the gross domestic product deflator index. “Use and intensity” includes quantity and mix of services. As a residual, this factor also includes any errors in measuring prices or total spending. “Age-sex mix” refers to that mix in the population. Growth in the total PHC Price Index is equal to the sum of economywide and relative PHC inflation and is a chain-weighted index of the price for all personal health care deflators. The height of the bars reflects the sum of factors that contribute positively to growth. In those cases where a factor may contribute growth of less than zero, the net total growth is reflected by the line and associated point estimate noted for each period.
March 2019 38:3 Health Affairs 7
Downloaded from HealthAffairs.org on February 20, 2019. Copyright Project HOPE—The People-to-People Health Foundation, Inc. For personal use only. All rights reserved. Reuse permissions at HealthAffairs.org.
rollmentgrowthisalsoanticipatedtoslowgradually during these years, from 2.8 percent in 2020to2.1percentby2027—aratemoreconsistent with the pre-baby-boom period. By the end of the projection period the Medicare share of total health spending is projected to rise to 24.1 percentby 2027 from 20.2 percentin 2017. Medicaid Medicaid spending growth is expected to have been just 2.2 percent in 2018, down from 2.9 percent growth in 2017 (exhibit 1)—the fourth consecutive year of slowing growth following the ACA’s expansion of Medicaid coverage in 2014. The expected trend in2018,asinprioryears,isprincipallyexplained by slower growth in enrollment, which is projected to have slowed to 1.1 percent in 2018 from 2.0 percent the previous year (exhibit 2). While growth for nearly all Medicaid services is expected to have slowed in 2018, growth in the net cost for Medicaid managed care plans is expected to have rebounded, compared to a decline in growth in 2017. This pattern reflects the historical and projected timeline over which the federal government is recovering payments from managed care organizations as a result of favorable prior-period experience.1 Growth in Medicaid spending is expected to acceleratein2019to4.8percent.Fiveadditional states have approved and are expected to implement Medicaid expansion in 2019, a factor that contributes in part to the aggregate spending growthincrease.ProjectedMedicaidenrollment growth—2.4percentin2019comparedto1.1percentin2018—reflectsthisnewlyeligiblepopulation. Growth in per enrollee Medicaid spending is expected to accelerate, as well, by 1.3 percentage points to 2.4 percent in 2019, as a result of faster growth in price factors. Medicaid spending is expected to grow at an average rate of 6.0 percent over 2020–27. The patterninannualgrowth,however,isinfluenced byreductionstodisproportionatesharehospital payments for hospitals set in law.9 These payments are scheduled to be reduced in 2020 and are then further reduced in 2021. Consequently, Medicaid spending growth is expected to grow slowly at 5.0 percent in 2020 and 5.4 percent in 2021. For 2022 through 2025, when the disproportionate share hospital payment reductions are equivalent to 2021, overall Medicaid spending growth is expected to be higher at 6.1 percent. Beginning in 2026 there are no reductions in the disproportionate share hospital payments,whichleadstoanotableexpectedoneyear acceleration in 2026 for overall Medicaid spending growth to 7.0 percent. Otherwise, an enrollment mix more heavily influenced by spendingpatternsofcomparativelymoreexpensive aged and disabled beneficiaries is expected
to result in per enrollee spending growth that is at or above 5 percent in every year during 2022–27. Private Health Insurance And Out-OfPocket Spending For private health insurance spending, growth is expected to have increased slightlyfrom4.2percentin2017to4.5percentin 2018, near the overall growth rate for national health expenditures of 4.4 percent (exhibit 1). While spending for most services and goods is expected to have grown slightly faster in 2018,10 the acceleration was partially offset by slower projectedgrowthinthenetcostofprivatehealth insurance,11 as private insurers offering plans in the Marketplace had fared better financially in 2017 and thus reduced the difference between premium revenues and expected benefit payments.12 Out-of-pocket spending growth is expected to have accelerated to 3.6 percent in 2018 from 2.6 percent in 2017, a rate that is consistent with faster income growth as well as with the higher average deductibles for employer-based private health insurance enrollees in 2018 compared to 2017.13 The projected spending trends in 2019 in part reflect the estimated impact of the effective repeal of the individual mandate. As some people choose to forgo maintaining health insurance, private health insurance enrollment is expected to decline slightly, primarily in the directpurchaseinsurancemarket.Accordingly,private health insurance spending growth is expected to slow to 3.3 percent in 2019 from 4.5 percent in 2018. Conversely, out-of-pocket spending is expected to grow more rapidly,at 4.8 percentin 2019 compared to 3.6 percent in 2018, in part because fewer people have private insurance coverage. Private health insurance spending is expected to grow 5.1 percent per year, on average, for 2020–27. Growth in this spending is projected to peak at 5.4 percent in 2023–24, in lagged response to the high anticipated growth in disposable personal income a few years prior. Private health insurance spending growth is then expected to slow to 4.8 percent by 2027, as income growthgenerallydecelerates.As the payer with the slowest expected growth over the full projection period, the private health insurance shareofnationalhealthspendingisprojectedto fall from 33.9 percent in 2017 to 31.8 percent in 2027. Growth in out-of-pocket spending, which is also primarily influenced by economic factors, isexpectedtobesimilartothatofprivatehealth insurance spending in 2020–27, at 5.0 percent. However, the projection-period peak in growth is expected in 2022 (5.4 percent), the year in which the excise tax on high-cost insurance
Costs & Spending
8 Health Affairs March 2019 38:3
Downloaded from HealthAffairs.org on February 20, 2019. Copyright Project HOPE—The People-to-People Health Foundation, Inc. For personal use only. All rights reserved. Reuse permissions at HealthAffairs.org.
plans is scheduled to go into effect.14 By 2027, becausetotalout-of-pocketspendingisexpected to grow more slowly, on average, than health insurance spending (exhibit 1), it is expected to account for a decreasing share of national health spending (9.8 percent in 2027, down from 10.5 percent in 2017).
Outlook For Major Medical Services And Goods Prescription Drugs Following growth of just 0.4 percent in 2017, prescription drug spending isexpectedtohavegrown3.3percentin2018but still be among the slowest-growing health care sectors (exhibit 3). Higher utilization growth is anticipated, compared to the relatively low growth in 2016 and 2017,1 partially driven by anincreaseinthenumberofnewdrugintroductions (fifty-nine in 2018, up from an average of thirty-four during 2016–17).15 In 2019 prescription drug spending growth is projectedtoacceleratefurther,to4.6percent,as a result of higher expected growth in drug utilization(includingfromnewdrugs)andamodest increase in drug price growth. Prescription drug spending is expected to increase, on average, by 6.1 percent per year for 2020–27 (exhibit 3). Contributing to the acceleration in growth during this period is the expectation that the use of prescription drugs will increaseoverthenextseveralyearsasaresultof increasingly robust efforts by employers and insurerstoreduceanybarriersregardingtheuseof maintenance drugs needed to keep their enrolleeswithchronicconditionshealthy.16 Twoother factors contributing to higher expected growth in the use of prescription drugs are the aging of thepopulationandchangestopharmacotherapy guidelines.16 These trends, coupled with faster expected spending increases in lagged response to faster growth in income, result in a peak projected growth rate for prescription drug spending of 6.4 percent in 2023–24. Finally, prescription drug spending growth is expected to rise becauseofashiftintheintensityandmixofdrug usageassociatedwiththemanyprojectscurrently in clinical development that could, over the nextfewyears,resultininnovative,yetmoreexpensive,newdrugsacrosssuchtherapeuticareas as cancer, diabetes, and Alzheimer’s disease.17 Hospitals Hospital spending is expected to have grown similarly in 2018 (4.4 percent) and 2017 (4.6 percent) (exhibit 3). By payer, somewhatslowergrowthinbothMedicaidandprivate healthinsurancehospitalspendingoffsetslightly faster growth in Medicare hospital spending. For2019hospitalspendinggrowthisexpectedto increase to 5.1 percent because of faster growth
in Medicare hospital payment updates and an increaseintheuseofhospitalservicesassociated withnewMedicaidexpansion–relatedenrollees. These increases are somewhat offset by slower expectedgrowthinprivatehealthinsurancehospital spending, which is partially attributable to the repeal of the individual mandate. Over2020–27hospitalspendinggrowthisexpected to average 5.7 percent per year, up from 5.1 percent in 2019. Consistent with overall spending, Medicare is expected to experience the fastest growth in spending for hospital care during this period. The peak growth for overall hospital spending is projected to occur in 2026 (6.1 percent) and is strongly influenced by substantially faster Medicaid spending growth in 2026 that reflects the expiration of Medicaid disproportionate share hospital payment reductions scheduled in current law for September 30, 2025. Private health insurance spending growth for hospital care is expected toreachitsprojection-periodpeakin2024,consistent with the lagged relationship to income. Hospital price growth is also expected to rise by2027.Theaccelerationinthisgrowthoverthe projection period primarily reflects continued wage increases for hospital employees that are anticipated from the low rates of growth experienced following the Great Recession, as well as tighter labor markets for hospital employees, including nurses.18 Growth is partially offset, however, by Medicare payment updates that are reduced by growth in economywide productivity,whichisprojectedtoaccelerateduringthe projection period.4 Physician And Clinical Services Spending in2018forphysicianandclinicalservicesisprojected to have grown 4.9 percent, rising from 4.2 percent in 2017 (exhibit 3). Price growth for physician and clinical services is expected to have increased 0.3 percentage point but to have remained at near historically low rates at 0.7percent.Thiscontinuedlowpricegrowthwas likely influenced, in part, by physician practices using more nonphysicians to provide care, a practicethatwasrelatedtoincreasedproductivity and profits even in the presence of slow price growth.8 The acceleration in overall projected spending growth also reflects faster growth in use that is partly related to a lagged response to growth in income over the recent history and alsofromincreasesinthenumberofofficevisits due to the severe 2017–18 flu season.19 In2019,growthinspendingforphysicianand clinical services is projected to accelerate once more,to5.4percentfrom4.9percentin2018.An accelerationinMedicaidspendinggrowthisthe primary factor contributing to the trend, which is in part associated with program’s expansion
March 2019 38:3 Health Affairs 9
Downloaded from HealthAffairs.org on February 20, 2019. Copyright Project HOPE—The People-to-People Health Foundation, Inc. For personal use only. All rights reserved. Reuse permissions at HealthAffairs.org.
by additional states. Over the remainder of the projection period, 2020–27, average annual growth in physician and clinical services spending is projected to be 5.4 percent. The growth rate for Medicare spending is expected to be substantially faster than that projected for physician and clinical services spending in private health insurance. Thatprojecteddifferentialislargelyduetofaster enrollment associated with the continued shift ofthebaby-boomgenerationfromprivatehealth insurance to Medicare. Another factor contributing to the growth in overall physician and clinical services spending over 2020–27 is an anticipated acceleration in physicianprice growth.Underlyingthisacceleration are projected rising costs related to the provision of care. In particular, wages are expected to increase as a result of the supply of physicians not being able to meet expected increases in demand for care connected with the aging population.20 Furthermore, some of the productivity gains that have been achieved throughtheuseoflower-costprovidersasasubstitute for physician care within physician practices may be less pronounced in the future, becauseoflimitationssuchaslicensingrestrictions on the scope of care that may be provided by nonphysician providers.21
Conclusion Duringthepasttenyearsthelingeringeffectsof the Great Recession, coupled with the coverage
and payment provisions of the Affordable Care Act, have significantly influenced the trends in healthcarespendingandenrollmentintheUnited States. Over the next decade, however, the outlook for health spending and insurance coverage isexpected to beprimarily drivenby longobserved demographic and economic factors fundamental to the health sector. While the national health spending growth rate is projected to average 5.5 percent per year for 2018–27 (exhibit 1), annual growth is expected to generally accelerate over much of the projection period. Medicare spending growth is expected to accelerate and be the fastest among the major payers, reflecting not only the continued enrollment shift of the baby-boom generationintotheprogrambutalsothegrowthratefor useandintensity,whichisprojectedtogradually increase toward the rates observed during Medicare’s long-term history. Growth in health care prices,reflectingbotheconomywideandrelative personal health care price inflation, is also expected to rebound somewhat toward rates more consistent with the period before the Great Recession and to return to a state in which personal health care price growth exceeds that of economywidepriceinflation.Finally,recentand anticipatedfastergrowthindisposablepersonal income is expected to lead to an increased demand for services, albeit with a lag, and put upwardpressureonthepatternofprivatehealth insurance and out-of-pocket spending growth over the projection period. ▪
The opinions expressed here are the authors’ and not necessarily those of the Centers for Medicare and Medicaid
Services. The authors thank Paul Spitalnic, Stephen Heffler, Aaron Catlin, Micah Hartman, Greg Savord, Cathy
Curtis, and anonymous peer reviewers for their helpful comments. [Published online February 20, 2019. ]
NOTES
1 Martin AB, Hartman M,Washington B, Catlin A, National Health Expenditure Accounts Team. National healthcarespendingin2017:growth slows to post–Great Recession rates; share of GDP stabilizes. Health Aff (Millwood). 2019;38(1):96–106. 2 By 2019 the individual mandate repeal is anticipated to result in about 1.5 million fewer direct-purchasemarket enrollees, who are expected to be somewhat younger and healthier than those who retain coverage, aswellasabout1.0 million fewer employer-sponsoredinsurance-market enrollees, than otherwise would have been projected. After 2019 the enrollment effects are expected to be smaller. Medicaid enrollment is assumed to be unaffected. See Centers for Medicare and Medicaid Services.
Projections of national health expenditures (note 3). 3 Centers for Medicare and Medicaid Services. Projections of national health expenditures: methodology and model specification [Internet]. Baltimore (MD): CMS; 2018 Feb 14 [cited 2019 Feb 4]. Available from: https://www.cms.gov/ResearchStatistics-Data-and-Systems/ Statistics-Trends-and-Reports/ NationalHealthExpendData/ Downloads/Projections Methodology.pdf 4 Boards of Trustees. 2018 annual report of the Boards of Trustees of the Federal Hospital Insurance and Federal Supplementary Medical Insurance Trust Funds [Internet]. Baltimore (MD): Centers for Medicare and Medicaid Services; 2018 [cited 2019 Jan 25]. Available from:
https://www.cms.gov/ResearchStatistics-Data-and-Systems/ Statistics-Trends-and-Reports/ ReportsTrustFunds/Downloads/ TR2018.pdf 5 Consistent with the methods employed in the Medicare Trustees Report (see note 4), these projections assume that payments would continue to be made even after the projected depletion of the Medicare Hospital Insurance trust fund, currently projected to occur in 2026. 6 Personal health care expenditures (PHC) measures the total amount spent to treat people with specific medical conditions. It represents about 85 percent of total national health expenditures over the projection period. 7 Brot-Goldberg ZC, Chandra A, Handel BR, Kolstad JT.What does a
Costs & Spending
10 Health Affairs March 2019 38:3
Downloaded from HealthAffairs.org on February 20, 2019. Copyright Project HOPE—The People-to-People Health Foundation, Inc. For personal use only. All rights reserved. Reuse permissions at HealthAffairs.org.
deductible do? The impact of costsharing on health care prices, quantities, and spending dynamics [Internet]. Cambridge (MA): National Bureau of Economic Research; 2015 Oct [cited 2019 Jan 25]. (NBER Working Paper No. 21632).Availablefrom:https://www .nber.org/papers/w21632.pdf 8 Medical Group Management Association [Internet]. Englewood (CO): MGMA; 2018. Press release, New MGMA data shows medical practices utilizing more non-physician providers are more profitable, productive; 2018 Jul [cited 2019 Jan 25]. Available from: https://www.mgma .com/news-insights/press/newmgma-data-shows-medicalpractices-utilizing-mo 9 The current schedule of reductions to Medicaid disproportionate share hospital payments was most recently modified by the Bipartisan Budget Act of 2018. 10 Mercer. Mercer annual survey finds health benefit cost growth will hold at4.1%in2019[Internet].NewYork (NY): Mercer; 2018 Sep 12 [cited 2019 Jan 25]. Available from: https://www.mercer.com/ newsroom/mercer-annual-surveyfinds-health-benefit-cost-growthwill-hold-at-41-in-2019.html 11 The net cost of insurance is the difference between total private health insurance spending and benefits incurred. It includes administrative costs, taxes, net gains or losses to reserves, and profits.
12 Fehr R, Cox C, Levitt L. Individual insurance market performance in mid-2018 [Internet]. San Francisco (CA): Henry J. Kaiser Family Foundation; 2018 Oct 5 [cited 2019 Jan 25]. Available from: https://www .kff.org/health-reform/issue-brief/ individual-insurance-marketperformance-in-mid-2018/ 13 Claxton G, Rae M, Long M, Damico A, Whitmore H. Health benefits in 2018: modest growth in premiums, higher workercontributions at firms with more low-wage workers. Health Aff (Millwood). 2018;37(11): 1892–900. 14 Inresponsetotheexcisetaxonhighcost insurance plans, some employers are expected to reduce the value of their health insurance benefits to remain below tax thresholds, which would result in higher cost sharing for employees. 15 Food and Drug Administration, Center for Drug Evaluation and Research. Advancing health through innovation: 2018 new drug therapy approvals [Internet]. Silver Spring (MD):FDA;2019Jan[cited2019Jan 25]. Available from: https://www .fda.gov/downloads/Drugs/ DevelopmentApprovalProcess/ DrugInnovation/UCM629290.pdf 16 IQVIA. 2018 and beyond: outlook andturningpoints.Parsippany(NJ): IQVIA; 2018 Mar 13. 17 EvaluatePharma. World preview 2018, outlook to 2024 [Internet]. London: EvaluatePharma; 2018 Jun [cited 2019 Jan 25]. Available from:
http://info.evaluategroup.com/rs/ 607-YGS-364/images/WP2018.pdf 18 Evans M. U.S. hospital profits fall as labor costs grow and patient mix shifts.Wall Street Journal [serial on the Internet]. 2018 Apr 23 [cited 2019 Jan 25]. Available from: https://www.wsj.com/articles/u-shospital-profits-fall-as-labor-costsgrow-and-patient-mix-shifts1524495601 19 Garten R, Blanton L, Elal AIA, Alabi N, Barnes J, Biggerstaff M, et al. Update: influenza activity in the United States during the 2017–18 season and composition of the 2018–19 influenza vaccine. MMWR Morb Mortal Wkly Rep. 2018; 67(22):634–42. 20 Dall T, West T, Chakrabati R, Reynolds R, Iacobucci W. 2018 update: the complexities of physician supply and demand: projections from 2016 to 2030: final report [Internet].Washington (DC): IHS Markit; 2018 Mar [cited 2019 Jan 25]. Available from: https://aamcblack.global.ssl.fastly.net/ production/media/filer_public/85/ d7/85d7b689-f417-4ef0-97fbecc129836829/aamc_2018_ workforce_projections_update_ april_11_2018.pdf 21 Hoffman M. Can nurse practitioners fill the void in primary care? MD [serial on the Internet]. 2018 Apr 19 [cited 2019 Jan 25]. Available from: https://www.mdmag.com/medicalnews/can-nurse-practitioners-fillthe-void-in-primary-care
March 2019 38:3 Health A

Concept of a Love for Cash is introduced: Physical money worship & strong emotional attachment to it; next to fear of technology, is one of the most important reasons for the resistance to cashless banking

Banking, Unbanking, and New Banking. Dominika Maison. The Psychology of Financial Consumer Behavior pp 185-208, https://link.springer.com/chapter/10.1007/978-3-030-10570-9_6

Abstract: This chapter is dedicated to analysing the psychological factors that support or hinder different levels of banking (having an account, use of a payment card, and use of mobile banking). There are still many people in countries across the world who do not have a bank account or do not use payment cards, which leads to financial exclusion. Apart from the obvious reason for being unbanked resulting from infrastructure limitations (e.g., limited access to bank branches or payment terminals), there are also psychological factors that can affect the level of banking service use. A new model of the levels of banking service use and the results of quantitative and qualitative research are presented in this chapter, revealing the specific nature of functioning on each of the identified levels of banking, along with the psychological barriers relating to moving up banking levels. The barriers to having a bank account and to acceptance of cashless financial behaviours are discussed. A new concept of a Love for Cash is introduced that refers to physical money worship and strong emotional attachment to physical money. Love for Cash, next to fear of technology, was found to be one of the most important reasons for the resistance to cashless banking.

Keywords: Financial exclusion Model of levels of banking service use Love for Cash scale Money worship Cashless financial behaviour

Los Angeles county: Streetcars' influence, visible in current urban density, has not dissipated in the 60 years since the streetcar's removal; mutually reinforcing pathways of regulation & agglomerative clustering

Vestiges of Transit: Urban Persistence at a Micro Scale. Leah Brooks and Byron Lutz. Review of Economics and Statistics, March 04, 2019. https://doi.org/10.1162/rest_a_00817

Abstract: We document intra-city spatial persistence and its causes. Streetcars dominated urban transit in Los Angeles County from the 1890s to the early 1910s, and were off the road entirely by 1963. However, we find that streetcars' influence remains readily visible in the current pattern of urban density and that this influence has not dissipated in the 60 years since the streetcar's removal. We examine land use regulation both as a consequence of streetcars and as a mechanism for streetcars' persistent effect. Our evidence suggests that the streetcar influences modern behavior through the mutually reinforcing pathways of regulation and agglomerative clustering.

EconLit codes: R3, R4, R5, N9

 We document intra-city spatial persistence and its causes.

Bias Blind Spot: They rated themselves lower than others in both susceptibility to biases & personal shortcomings; belief in free will was positively associated with the susceptibility to better than average effect

Agency and self-other asymmetries in perceived bias and shortcomings: Replications of the Bias Blind Spot and extensions linking to free will beliefs. Prasad  Chandrashekar et al. March 2019. DOI: 10.13140/RG.2.2.19878.16961

Description: Bias Blind Spot is the phenomenon that people tend to perceive themselves to be less susceptible to biases than others. In three pre-registered experiments with samples from Hong Kong and the United States (overall N = 969), we replicated two experiments (Study 1-Survey 2 and Study 2) from Pronin, Lin, and Ross (2002), the first published demonstration of the effect. Participants rated themselves lower than others in both susceptibility to biases (mini meta-analysis: dz = - 1.00) and personal shortcoming (mini meta-analysis: dz = - 0.34). The self-other asymmetry of susceptibility for biases was larger than that of personal shortcomings (dz = - 0.43). Thus, the replication findings provide strong empirical support for the bias blind spot phenomenon. Extending the replications, belief in free will was positively associated with the susceptibility to better than average effect, and also with a stronger self-other asymmetry in rating personal shortcomings.


---
The bias blind spot is the phenomenon that people tend to perceive themselves as less
biased than others in their judgements and behaviors (Pronin, 2007; Pronin et al., 2002) and
further tend perceive their peers as more subject to the bias blind spot than they are. Broadly,
people seem to be able to detect or infer biases in others but fail to do so about themselves
(Pronin, Gilovich, & Ross, 2004). 
We had two main goals for the present investigation: (1) to conduct replications of the
bias blind spot effect, and (2) to examine extensions about the link between agency beliefs and
the bias blind spot. The proposed extension was meant to answer further calls for future research
to investigate possible “causes of people’s unwarranted faith in their own introspections?”
(Pronin, 2007, p. 41). We begin by introducing the literature on the bias blind spot and the chosen
target article for replication - Pronin et al. (2002). We then introduce agency beliefs and
hypothesize about the relationship between free will beliefs and the bias blind spot effect.
Bias blind spot and free will beliefs                  5
Bias blind spot
People have access to their own private mental lives but not to that of others, and so to
overcome this information asymmetry people continuously aim to detect and infer other people's
internal psyche. This information asymmetry often results in unintended and unaware self-other
attributional asymmetries (Nisbett & Ross, 1980). 
Many such attribution asymmetries have been documented over the years, such as the
widely discussed actor-observer bias (Jones & Nisbett, 1972), in that divergent responses of
others are perceived to reflect others’ stable personality dispositions (Jones, 1990). Individuals
also tend to believe and perceive their perceptions of the world are objective and accurate, coined
as "naïve realism" (Ross & Ward, 1995), easily detecting flaws in this assumption in others but
failing to detect such flaws in themselves. They may further assume that they self-reflect more
than others do, since when people evaluate their own behaviors and judgments, they base their
evaluations on introspection, whereas when evaluating others their assessments must rely on
others' behavior (Pronin, 2007; Pronin et al., 2004; Ross & Ward, 1995; Wilson & Dunn, 2004).
Finally, people tend to view themselves in a positive light or construe reality in a way that would
make for a more positive self-evaluations. Together, naïve realism, introspection illusion, and
self-enhancement motives, result in a bias blind spot, a self-other asymmetry in perceptions bias.
People further reject and persist in their biased perceptions even if being made aware of its
existence (Ehrlinger, Gilovich, & Ross, 2005; Pronin, 2007; Pronin & Kugler, 2007).
The chosen bias blind spot experiments
Pronin et al. (2002)’s work was the first demonstration of the bias blind spot. The article
has been influential with over 765 citations according to Google Scholar at the time of writing,
with theoretical developments and applications across several domains, such as judgment and
Bias blind spot and free will beliefs                  6
decision making, behavioral economics, assessment, and interpersonal and intergroup conflict
(e.g., Pronin et al., 2004). Bias blind sport is argued to be a distinct meta-bias with clear
implications on judgment and behavior (Scopelliti et al., 2015). For example, subsequent
empirical research showed relevance of the effect to law in viewing and remembering criminal
events (Jones, Crozier, & Strange, 2018), with observations of the effect in children as young as
six years old (Hagá, Olson, & Garcia-Marques, 2018), and evidence for the persistence of bias
despite being shown how it affected previous decisions (Hansen, Gerbasi, Todorov, Kruse, &
Pronin, 2014). To the best of our knowledge, there have been no previous attempts for a direct
replication of the experiments reported in the paper. 
The target article consisted of three experiments, and the current replications focused on
Survey 2 of Study 1 and Study 2. Findings in Pronin et al. (2002) and summarized in Table 1.
In their Study 1, participants separately rated their own and average Americans'
susceptibility to eight cognitive and motivational biases and three personal shortcoming biases.
Pronin et al. (2002) proposed that participants would demonstrate clear asymmetry in
susceptibility to biases but not to shortcoming items because people are likely to be aware of their
visible personal shortcomings than their invisible cognitive biases. In view of this, they proposed
and tested three hypotheses. The first hypothesis suggested that participant would rate themselves
as less susceptible than others on biases. The second hypothesis suggested that there should be no
difference between the rating of self-susceptibility and other’s susceptibility towards personal
shortcomings (null hypothesis). The third hypothesis combining the first two suggested that
asymmetries effects would, therefore, be larger for biases than for personal shortcoming items.
Bias blind spot and free will beliefs                  7
In their Study 2, participants compared themselves to others on six personality
dimensions, three positive and three negative. Findings indicated that participants rated
themselves as higher on positive personality dimensions and lower on negative ones, compared to
others. Furthermore, a large majority (76 %) of participants who claimed better-than-average
status insisted on this status even when alerted to the possible bias.
Adjustments to original design
We attempted a close replication of the original study by Pronin et al. (2002) yet made
several needed adjustment. First, we administered all surveys via an online Qualtrics survey.
Second, the two studies from Pronin et al. (2002) chosen for replication included Stanford
University undergraduate students and were not paid for completion of the study. The current
replication effort of three studies included one undergraduate sample from a university in HK,
and two paid samples using Amazon Mechanical Turk (MTurk). Thirdly, Study 2 and Study 3
replications combined the original article' Studies 1 and 2 into an integrated design. Forth, we
went beyond the replication and added extensions to the original design to examine the link
between agency beliefs and self-other bias asymmetries.
Extensions - belief in free will as a predictor of the bias blind spot
We aimed to extend the replication study by considering individuals’ perceived agency as
a predictor of the bias blind spot - whether perceiver’s beliefs in free will predict self-other
asymmetries regarding biases and personal shortcomings. Belief in free will is the general belief
that human behavior is free from internal and external constraints across situations (Feldman,
2017; Monroe & Malle, 2014). Studies on folk understanding of free will found that people
normally associate free will with having choice and understand free will as the absence of
Bias blind spot and free will beliefs                  8
internal and external constraints (Baumeister, 2008; Feldman, Baumeister,  & Wong, 2014;
Monroe, Dillon,  & Malle, 2014; Monroe & Malle, 2010; Vonasch, Baumeister, & Mele, 2018).
Belief in free will has been associated with a range of adaptive behavioral and
psychological outcomes such as academic and job performance (Feldman, Chandrashekar, &
Wong, 2016; Stillman et al., 2010), perseverance for long-term goals (Li, Zhao, Lin, Chen, &
Wang, 2018), self-control (Goto, Ishibashi, Kajimura, Oka,  & Kusumi, 2018), expressions of
love (Boudesseul, Lantian, Cova, & Bègue, 2016), job satisfaction (Feldman, Farh, & Wong,
2018), cooperation (Protzko, Ouimette, & Schooler, 2016), and well-being and meaning in life
(Crescioni, Baumeister, Ainsworth, Ent, & Lambert, 2016; Moynihan, Igou, & van Tilburg,
2017).
Although the research outlined above largely emphasize the effects of believing in free
will on individuals’ self-regulatory behavior, some recent work suggests that free-will beliefs
affect fundamental social-cognitive processes that are implicated in the judgements of self and
others. For example, Genschow, Rigoni, and Brass (2017) found a positive relationship between
the strength of the belief in free will and the correspondence bias, i.e., tendency to endorse
dispositional explanations than situational explanations. People with a strong belief in free will
tend to view the actions of wrongdoers as the result of choices that are freely made and therefore
endorse harsher punishments (Clark et al., 2014; Martin, Rigoni, & Vohs, 2017). Similarly, free
will beliefs influence judgements about the extent to which individuals’ choices determine their
economic outcomes (Mercier et al., 2018), and more broadly form the basis of a capacity for
change, not only for others but also for the self (Feldman, Wong, & Baumeister, 2016).
Bias blind spot and free will beliefs                  9
As discussed above, several findings are suggestive of the possible relationship between
free will beliefs and the bias blind spot. Personal shortcomings can be viewed as a form of
internal constraints of free will (e.g., fear of public speaking, planning fallacy, and
procrastination). Those with stronger free will beliefs are likely to perceive own behaviors as the
outcome of their own choices (Feldman et al., 2014) and have more control over decisions in life
(Rigoni, Kühn, Gaudino, Sartori, & Brass, 2012; Stillman, Baumeister, & Mele, 2011). In
summary, free will beliefs reflect a view of the self as an active agent with freedom to choose
actions and pursue goals, and therefore, should encourage the view of the self as devoid of
internal constraints that may impose limitations on self’s ability to make choices. Supporting this
view, people with a history of addiction to alcohol, tobacco, and other drugs rate themselves
lower on belief in free will (Vonasch, Clark, Lau, Vohs, & Baumeister, 2017). This is suggestive
of a negative relationship between free will beliefs and perceived personal shortcomings.
Furthermore, the work on the association between free will beliefs and correspondence bias
suggest that free-will beliefs would be associated with perceiving others as more affected by their
personal shortcomings.  Combined, the two are suggestive of a positive relationship between free
will beliefs and bias blind spot regarding personal shortcoming. We initially made no pre
registered predictions regarding perceived biases although similar arguments can be made.
Exploratory Hypotheses
We did not make priory predictions regarding associations between free will beliefs and
susceptibility to biases of the self and others. Belief in free will at its core is experienced as an
increased sense of agency, therefore believers perceive their own behaviors as generated by
themselves, rather than external forces (Rigoni et al., 2012). They therefore view their own
judgements and behaviors as lacking in biases, and by extension, may exhibit larger self-other
Bias blind spot and free will beliefs                  10
asymmetry in perceived bias. Extending on this argument toward susceptibility to better than
average effect, the belief in free will likely to be negatively associated with ratings of negative
personality dimensions of self in comparison to others, and positively associated with positive
personality dimensions of self in comparison to others.
Overview of empirical studies
There was a two weeks gap between the two data collections of Studies 1 and 2. In each of
the replication studies, we first pre-registered the experiment on the Open Science Framework
(OSF) and data collection was launched later that week. Pre-registrations, power analyses, and all
materials used in these experiments are available in the supplementary materials. OSF pre
registration review links: Study 1a -
https://osf.io/fwthk/?view_only=744526890b674a9fbec72acc37a79c86 ; Study 1b -
https://osf.io/qmcrn/?view_only=4820ad08078b4b5a860b08c0234c7229 ; Study 2 -
https://osf.io/fm48b/?view_only=60e6cf6df39147e0af1b28f4e7da0d4c.
In light of findings from the first two studies, Study 3 was designed to extend the findings.
Importantly, we wanted to replicate the proposed extensions in the Study 2 with a larger sample
to be able to detect smaller effect sizes. We preregistered our hypotheses and analysis plan on the
OSF, review link: https://osf.io/u3vds/?view_only=42450fc3d6b74866a1c022e7bfd299a9. 
Data and R/RMarkdown code for all studies is available on the OSF, review link:
https://osf.io/3df5s/?view_only=b29f8571eb874448907ce45c7379e371 . Full open-science
details and disclosures are provided in the supplementary. All measures, manipulations,
exclusions conducted for this investigation are reported, all studies were pre-registered with
Bias blind spot and free will beliefs                  11
power analyses reported in the supplementary, and data collection was completed before
analyses. 
Studies 1a and 1b
Studies 1a and 1b were meant as a pre-test of the effects in an undergraduate class.
Students worked in teams of 3-6 to design and run a series of replications, two of those were
Pronin et al.'s Study 1 Survey 2 and Study 2 corresponding to our Study 1a and 1b. The students
then served as the target sample for the experiments designed by their classmates, experiments
they were not involved in designing and had no prior knowledge of. The course materials covered
judgement and decision-making biases, which meant that the students were made aware a wide
array of other biases, and the experiments are, therefore, very conservative tests of the effect in a
non-naive sample.
Students were randomly assigned into groups and to the study for replication. Student
groups designed the survey, conducted effect size and confidence intervals calculations,
conducted power analyses, and wrote the pre-registrations for Studies 1a and 1b. The course
instructor completed the pre-registration on OSF and data collection. All the students registered
in the course were invited to take part as respondents in the study. To ensure anonymity, students
were only asked to indicate which replication group they belonged to and those were later
excluded from the data analysis of the study they designed. 
Participants and procedures
A total of 49 undergraduate students took part in the online course survey, and of those we
excluded the four students who designed Study 1a and six students who designed Study 1b,
Bias blind spot and free will beliefs                  12
resulting in a sample of 45 for Study 1a (Mage = 20.20, SD = 0.99; 31 females) and 43 for Study
1b. 
Study 1a
Measures
Biases and Personal shortcomings. 
Participants were presented with descriptions of eight biases and three personal
shortcomings: self-serving attributions for success or failures, dissonance reduction after free
choice, positive halo effect, biased assimilation of new information, reactive devaluation of
proposal one’s negotiation counterparts, perceptions of hostile media bias toward one’s group or
cause, fundamental attribution error (FAE) in “blaming the victim,” and judgments about the
“greater good” influences personal self-interest, procrastination, fear of public speaking, and
planning fallacy. The supplementary includes detailed descriptions of the biases and personal
shortcomings.
For each of the descriptions participants rated on their own susceptibility and
susceptibility of the average student at the university. Ratings were on a nine-point scale (1 = not
at all; 9 = strongly). 
Results and discussion
Descriptive statistics of the ratings on the susceptibility to bases biases and personal
shortcomings are presented in Table 2 (see supplementary for the descriptive statistics and plots
for each of the biases and personal shortcomings). We conducted the paired sample t-test to test
the hypothesis, summarized in Table 3.
Bias blind spot and free will beliefs                  13
Results of paired t-tests (one-tailed) indicated that participants, consistent with the original
study, reported themselves as less susceptible to biases (M = 5.60, SD = 0.86), than the average
students in the university (M = 6.35, SD = 0.91), Md  = -0.75, t (44) = -4.54, p <.001, dz = -0.68,
95% CI [-1.01, -0.35] (Plotted in Figure 1). Self-others asymmetry was found for all individual
biases except for cognitive dissonance (Table S3 in the supplementary).
In the original study, the authors made no prediction regarding self-other personal
shortcomings asymmetry. We conducted a two-tail dependent t-test but failed to find support for
any differences with a weak effect for high ratings of self (self: M = 6.20, SD = 1.78; others: M  =
6.49 , SD = 1.23; Md  = -0.29; t (44) = -1.13, p = .265; dz = -0.17, 95% CI [-0.47, 0.13]; see Figure
2 and Table S3 in the supplementary for details per each shortcoming). Quite possibly, as in the
original article, the small sample failed to detect a weak effect.
Finally, self-other bias asymmetry (M = -0.75, SD = 1.11) was stronger than self-other
personal shortcomings asymmetry (M = -0.29, SD = 1.72; Md = -0.46, t (44) = -1.97, p = .055, dz
= -0.29, 95% CI [-0.60, 0.01]; see Figure 3).
Study 1b 
Measures
Assessed Personality dimensions. 
Participants were presented with three positive and three negative personality dimensions
in randomized order. The positive personalities assessed were dependability, objectivity, and
consideration. The negative personalities assessed were snobbery, deceptiveness, and selfishness.
The ratings were made on a 9-points (1 = much lower than the average student; 5 = same as the
average student; 9 = much higher than the average student).
Bias blind spot and free will beliefs                  14
Bias recognition. 
After rating their personalities, participants were briefed of the better-than-average effect
and asked whether they were influenced by the bias when assessing their personalities (1  -
Objective measures would rate me lower on positive characteristics and higher on negative
characteristics than I rated myself; 2 - Objective measures would rate me neither more positively
nor more negatively than I rated myself; 3 - Objective measures would rate me higher on positive
characteristics and lower on negative characteristics than I rated myself). 
Results and discussion
Table 4 details descriptive statistics and Table 5 summarizes statistical tests (Table S4 in
the supplementary details ratings for each personality dimension).
We conducted one-sample one-tail t-tests and found that participants rated themselves as
having more positive personality dimensions (M = 5.74, t (42) = 5.09, p < .001, dz = 0.78, 95% CI
[0.43, 1.11]) and less negative personality dimensions than others (M = 4.16, t (42) = -4.55, p <
.001, dz = -0.69, 95% CI [-1.02, -0.36]). 
We then conducted a chi-squared test to test the hypothesis that the majority of
participants deny having the better-than-average effect, comparing to a 50%-50% random split.
Despite being made aware of the potential bias, only 9 of the 43 participants (21%)
acknowledged their potential bias leaving 79% of participants still claiming to be better than their
average peers (χ2 (1, N = 43) = 14.53, p < .001, dz = 1.43, 95% CI [0.69, 2.16]). 
Findings supported the better than average effect and denial of their own bias. Effect size
(dz = 1.43) of the replication was almost two times greater than the effect size of the original
Bias blind spot and free will beliefs                  15
study (dz  = 0.70) and the replication’s confidence intervals ([0.69, 2.16]) includes the original
effect size point estimate. We conclude the replications as successful.
Study 2
Method
Participants and procedures
A total of 303 American Amazon Mechanical Turk (MTurk) participants completed the
study using TurkPrime.com (Mage = 38.45, SD = 11.58; 166 females). First, participants rated
their free will beliefs on two scales and then rated their and others susceptibility to the
descriptions of eight biases and three personal shortcomings. The design was a 2 (self and other
ratings) by 2 (biases and personal shortcomings) within-subject design and display of conditions
was counterbalanced (see supplementary for more details and full measures). Participants then
answered a funneling section and provided demographic information. 
Measures
Belief in Free will. 
Free will beliefs (BFW) were measured using two free-will belief subscales: 5 items
measure of general BFW (Nadelhoffer, Shepard, Nahmias, Sripada, & Ross, 2014) (1 = Strongly
disagree, 7 = Strongly agree; 𝛼 = 0.91) and BFW personal agency subscale (Rakos, Laurene,
Skala, & Slane, 2008) (4 items; 1 = Not true at all, 5 = Almost always true; 𝛼 = 0.92). Details of
all measures are provided in the supplementary. 
Bias blind spot and free will beliefs                  16
Biases and Personal shortcomings. 
Similarly to Study 1a, participants rated their own and other average Americans
susceptibility to biases and personal shortcomings (1 = not at all; 9 = strongly).
Results
Descriptive statistics are provided in Table 6 and statistical tests summary in Table 7 (see
Table S6 and Table S7 and Figures S5 to S8 in the supplementary for each of the biases and
personal shortcomings separately).We conducted a dependent sample t-test and found that
participants' perceived susceptibility to biases for self (N = 303; M = 4.64, SD = 1.35) was lower
than of others (M = 5.78, SD = 1.16 ; Md = -1.15; t (302) = -16.16, p < .001; dz = -0.93, 95% CI [
1.06, -0.79]; see Figure 4), and the self-other asymmetry effects were similar across all eight
biases (p < .001; see Table S8 in supplementary). In comparison, the original study found support
for only four of the eight biases and with weaker effects, possibly due to lacking power.
The original study found no support for self-other asymmetry in perceived personal
shortcomings, and the hypothesis was for a null (or weaker) effect. We conducted a dependent
sample t-test and found perceived personal shortcomings (M = 5.35, SD = 1.88) were lower than
perceived susceptibility to biases of other MTurk workers (M = 5.87, SD = 1.35; Md  = -0.52; t
(302) = -5.22, p <.001, dz = -0.30, 95% CI [-0.42, -0.18]; see Figure 5). The original study
reported self as lower than others for all three perceived personal shortcomings, yet it was not
reported if any of the results reached statistical significance. Our dependent sample t-tests found
support for an asymmetry for two out of three personal shortcomings (procrastination and
planning fallacy; see Table S8 in supplementary for details on each of the personal
shortcomings). These findings deviate from the findings of the original study.
Bias blind spot and free will beliefs                  17
Based on the findings in the original study, we expected a significant difference between
the biases and personal shortcomings asymmetries. We conducted a dependent sample t-test and
indeed found that self-other biases asymmetry (M = -1.15, SD = 1.24) was larger than the self
other personal shortcomings asymmetries (M = -0.52, SD = 1.75; N = 303 ; Md = -0.62, t (302) =
6.39, p <.001, dz = -0.37, 95% CI [-0.48, -0.25]; see Figure 6).
Finally, we examined the link between free will beliefs and perceived personal
shortcomings of self and others. Pearson correlations are detailed in Table 8. Both free will
beliefs scales were negatively correlated with perceived self personal shortcomings (general free
will: r = -0.22, p < .001, 95% CI [-0.32, -0.11]; personal agency: r = -0.17, p = .003, 95% CI [
0.28, -0.06]). However, we found no support for a link between free will beliefs measures and
perceived shortcomings in others (general free will: r = 0.00, p = .941, 95% CI [-0.11, 0.12];
personal agency: r = 0.05, p = .357, 95% CI [-0.06, 0.16]). Free will beliefs negatively correlated
with personal shortcomings self-other asymmetry (general free will: r = -0.24, p < .001, 95% CI
[-0.34, -0.13]; personal agency: r = -0.22, p < .001, 95% CI [-0.33, -0.11]). 
Probing the link between free will beliefs and susceptibility to biases we only found
support for personal agency subscale as negatively correlated with self-other asymmetry for
susceptibility to bias (r = -0.17, p  = .003, 95% CI [-0.28, -0.06]). We did not find support for
correlations between free-will beliefs and any of the measures associated susceptibility to biases:
self-bias (general free will: r  = -0.01, p = .811, 95% CI [-0.13, 0.10]; personal agency: r = -0.11, 
p = .055, 95% CI [-0.22, 0.00]), and others' bias (general free will: r = 0.02, p = .737, 95% CI [
0.09, 0.13]; personal agency: r  = 0.05, p = .366,  95% CI [-0.06, 0.16]).
Bias blind spot and free will beliefs                  18
Study 3
Method
Participants and procedures
A total of 621 American MTurk participants completed the study using TurkPrime.com
(Mage = 39.15, SDage = 11.88; 346 females). The rationale of the present study was as follows:
First, to test the robustness and reliability of the replication results found in Study 1a, Study 1b,
and Study 2 with a larger sample. Second, more importantly, replicate the proposed extension
hypotheses between free will beliefs and perceived personal shortcomings of self and others. 
Study 3 combined Study 1a and Study 1b as one single study. Procedures were modeled to
remain as close as possible to the original studies. The study included three parts. Participants
first completed the measures of free will beliefs. Then, in randomized order, participants rated
their and others susceptibility to given descriptions of eight biases and three personal
shortcomings, and compared themselves to others on three positives and three negative
personality dimensions with a test for recognition of their bias.
Measures
The measures for biases and personal shortcomings followed the design of Study 2. The
measures of personality dimensions and bias recognition followed were the same as Study 1b.
Free will beliefs were measured by three of the most common scales: eight items of the free-will
and determinism personal will sub-scales (Rakos et al., 2008) (0 = Not true at all, 4 = Almost
always true; 𝛼  = 0.74), five items of the general free will beliefs scale (Nadelhoffer et al., 2014)
(1 = Strongly disagree, 7 = Strongly agree; 𝛼 = 0.89) and the seven items from free will and
Bias blind spot and free will beliefs                  19
determinism plus scale (Paulhus & Carey, 2011) (1 = Not at all true, 5 = Always true; 𝛼  = 0.85;
recoded from a scale of 0 to 4 to match the original scale range).
Results
Table 9 details descriptive statistics and Table 10 summarizes statistical tests (see Table
S11 and Table S12 and Figures S9 to S12 in supplementary for each of the biases and personal
shortcomings separately).
Perceived susceptibility to bias and personal shortcomings
We conducted a series of dependent sample t-test mirroring Studies 1a and 2 (N = 621).
Perceived susceptibility to biases was lower for self (M = 4.69, SD = 1.30) than for others (M =
6.48, SD = 1.04; Md  = -1.80; t (620) = -32.04, p <.001, dz = -1.29, 95% CI [-1.39, -1.18]; see
Figure 7), in all biases (p < .001; see Table S13 in supplementary). Perceived shortcomings were
lower for self (M = 5.52, SD = 1.71) than for others (M = 6.25, SD = 1.16; Md = -0.73; t (620) =
10.54, p <.001, dz = -0.42, 95% CI [-0.51, -0.34]; see Figure 8), especially in procrastination and
planning fallacy (see Table S13 in supplementary). Self-other asymmetry was larger for biases
(M = -1.80, SD = 1.40) than for personal shortcomings (M = -0.73, SD = 1.73; Md  = -1.06; t (620)
= -13.01, p <.001, dz = -0.52, 95% CI [-0.61, -0.44]; see Figure 9). The results align with the
findings in Study 2.
Denying Personal Susceptibility to the Better Than Average Effect
We conducted a series of one-sample t-tests mirroring Study 1b. Participants rated
themselves as possessing more positive personality dimensions (M = 6.42; t (620) = 31.74, p
<.001; dz = 1.27, 95% CI [1.17, 1.38]) and less negative personality dimensions (M = 3.21; t
(620) = -30.38, p <.001; dz = -1.22, 95% CI [-1.32, -1.11]), compared to others. 
Bias blind spot and free will beliefs                  20
To assess denial of the bias, we conducted a chi-square comparing to a 50%-50% split.
Only 109 of the 621 participants (18%) admitted bias, leaving 82% denying the bias (χ2 (1, N =
621) = 261.53, p < .001; dz = 1.71, 95% CI [1.50, 1.91]).
Free-will beliefs and biases
Finally, we examined the link between free will beliefs and perceived personal
shortcomings of self and others. Pearson correlations are detailed in Table 11.
Belief in free will and personal shortcomings. 
Personal shortcomings for self was negatively associated with free-will beliefs (general: r
= -0.16, p < .001, 95% CI [-0.23, -0.08]; personal agency: r = -0.15, p < .001, 95% CI [-0.22,
0.07]; personal will: r = -0.09, p = .022, 95% CI [-0.17, -0.01]). However, we found no consistent
support for a link between free will beliefs measures and perceived shortcomings in others
(general: r = 0.02, p = .551, 95% CI [-0.05, 0.10]; personal agency: r = 0.05, p = .231, 95% CI [
0.03, 0.13]; personal will: r = 0.11, p = .007, 95% CI [0.03, 0.18]). 
Free will beliefs negatively correlated with personal shortcomings self-other asymmetry
(general free will: r = -0.17, p < .001, 95% CI [-0.25, -0.09]; personal agency: r = -0.18, p < .001,
95% CI [-0.25, -0.10]; personal will: r = -0.16, p < .001, 95% CI [-0.24, -0.09]).
Probing the link between free will belief and susceptibility to biases we found support for
personal will as negatively correlated with susceptibility to bias of the self (r = -0.14, p < .001,
95% CI [-0.22, -0.06]) and positively correlated with the susceptibility to bias of others (r = 0.12,
p = .003, 95% CI [0.04, 0.20]). Overall personal will was negatively correlated with self-other
asymmetry for susceptibility to bias (r = -0.22, p < .001, 95% CI [-0.29, -0.14]). We found no
Bias blind spot and free will beliefs                  21
support for a correlation with the two other measures of free will beliefs (correlations ranged
between 0.01 CI [-0.07, 0.09] and -0.03 CI [-0.11, 0.05]).
Belief in free will and better than average effect 
We found support for an exploratory negative relationship between free-will beliefs and
negative personality dimensions (general free-will: r = -0.09, p = .033, 95% CI [-0.16, -0.01];
personal agency: r = -0.16, p < .001, 95% CI [-0.23, -0.08]; personal will: r = -0.12, p = .003,
95% CI [-0.20, -0.04]). Positive personality dimensions were positively correlated with personal
will (r = 0.15, p < .001, 95% CI [0.07, 0.23]), but no support for a positive correlation with the
two other measures (General free will: r = 0.04, p = .341, 95% CI [-0.04, 0.12]; Personal agency:
r = 0.05, p = .222, 95% CI [-0.03, 0.13]).
Denial of bias correlated with general free-will (r = 0.11, p = .007, 95% CI [0.03, 0.19])
and personal agency (r = 0.14, p < .001, 95% CI [0.06, 0.22]), with no support for personal will
(r = 0.03, p = .531, 95% CI [-0.05, 0.10]). 
Overall, the results of the free will related findings are consistent with the results of Study
2.
General Results: Mini Meta-Analysis
We summarized the findings of the three studies together with the findings from original
article using a mini meta-analysis to assess the overall effect size (Goh, Hall, & Rosenthal, 2016;
Lakens & Etz, 2017). The overall effects for Study 1 Survey 2 of the original study were as
follows: bias asymmetry = -0.98 (95% CI = [-1.25, -0.72], p < .001) (see Figure  10), personal
Bias blind spot and free will beliefs                  22
shortcomings asymmetry = -0.19 (95% CI = [-0.47, 0.08], p = .158) (see Figure 11), bias versus
shortcomings difference = -0.44 (95% CI = [-0.56, -0.32], p < .001) (see Figure 12).
Similarly, overall effects for Study 2 of the original study were as follows: better than
average effect for positive personality dimensions = 1.22 (95% CI = [0.78, 1.66], p < .001) (see
Fig. 13), better than average effect for negative personality dimensions = -1.07 (95% CI = [-1.39,
-0.75], p < .001) (see Fig. 14), and denial to better than average effect = 1.32 (95% CI = [0.72,
1.91], p < .001) (see Fig. 15).
General Discussion
Summary and evaluation of replications
We conducted three replication studies of two studies from Pronin et al. (2002), testing the
bias blind spot effect. We summarized the findings of the three replication studies in Table 12.
Overall, we found that: (1) participants' perceived their susceptibility to biases as lower than that
of others, (2) participants perceived their own personal shortcomings as lower than that of others,
(3) bias asymmetry was larger than personal shortcomings asymmetry, (4) participants rated
themselves as higher on positive personality dimensions and lower on negative personality
dimensions, and (5) denied exhibiting the bias. 
The first aim of the current replication effort is to evaluate—in a confirmatory manner—
the size of an effect observed in the original study. To interpret the replication results we
followed the framework by LeBel, McCarthy, Earp, Elson, and Vanpaemel (2018) that take into
account three distinct statistical aspects of the results: (a) whether a signal was detected in the
replication (i.e., the confidence interval for the replication Effect size (ES) excludes zero), (b)
Bias blind spot and free will beliefs                  23
consistency of the replication ES with the original study’s ES, and (c) precision of the
replication’s ES estimate. The replication ES for asymmetry in bias in three individual studies
ranged between dz = -0.68 [-1.01, -0.35] and dz = -1.29 [-1.39, -1.18]. When pooled across all
studies with a mini-meta analysis, the overall estimate of the ES was: dz = -1.00 [-1.33, -0.67].
The results indicate signal was detected and that the replication ES is consistent with the original
study, i.e., the replication’s confidence interval includes the original ES point estimate of 0.86.
The replication results testing the asymmetry in personal shortcomings across three studies
ranged between dz = -0.17 [-0.47, 0.13] and dz = -0.42 [-0.51, -0.34]. Comparing the meta
analytic estimate (dz = -0.34 [-0.46, -0.23]) with original study suggest that, although, a signal
was detected, the replication ES is inconsistent and opposite in direction with the original ES
point estimate of 0.28. Therefore, a less favorable replication outcome suggesting small sample
size in the original study may have contributed to the observed effect. Finally, the hypothesis
testing the asymmetry between bias and personal shortcomings in current replication studies, ES
ranged between dz = -0.29 [-0.60, 0.01] and dz = -0.52 [-0.61, -0.44]. Meta-analytic estimate of
the ES (dz = -0.43 [-0.56, -0.29]) is inconsistent with the original ES point estimate of -0.61, i.e.,
similar in direction but smaller than the ES of the original study.
We followed a similar approach to summarize the replication of Study 2 of Pronin et al.
(2002). The replication ES’s for better than average effect for positive personality dimensions
were dz = 0.78 [0.43, 1.11] (Study 1b) and dz = 1.27 [1.17, 1.38] (Study 3). Meta-analytic
estimate of the ES (dz = 1.05 [0.57, 1.54]) is inconsistent with the original study’s ES point
estimate of 1.61, i.e., similar in direction but smaller than the ES of the original study. The
replication ES’s for better than average effect for negative personality dimensions were dz = -0.69
[-1.02, -0.36] (Study 1b) and dz = -1.22 [-1.32, -1.11] (Study 3). Meta-analytic estimate of the ES
Bias blind spot and free will beliefs                  24
(dz = -0.98 [-1.49, -0.47]) is consistent with the original study’s ES point estimate of -1.24.
Similarly, the replication ES for denial of bias were dz = 1.43 [0.69, 2.16] (Study 1b) and dz =
1.71 [1.50, 1.91] (Study 3). Meta-analytic estimate of the ES (dz = 1.69 [1.49, 1.88]) is
inconsistent with the original study’s ES point estimate of 0.76, i.e., similar in direction but larger
than the ES of the original study. 
In summary, the replication results show that ESs are similar in direction with the original
study for all the hypothesis tested and indicated signal (i.e., ES excludes zero) except for the
prediction of asymmetry in perceived shortcomings. As noted above effect size estimates in some
cases were inconsistent with the original study. However, we note that the sample size employed
in the original study was small. Overall, the replication results provide reasonable support for the
findings of the original study.
Agency beliefs extension
In Studies 2 and 3 we ran extensions examining the link between free will beliefs and the
bias blind spot effects, and the findings are summarized in Table 13. 
To this end, we pre-registered the theoretical relationship between the strength of belief in
free will on individuals’ tendency toward bias blindness and better than average effect. Overall,
our findings provide support for the hypothesis that belief in free will is linked asymmetry in
perceived personal shortcomings of self and of others. This particular asymmetry is mainly
driven by the negative correlation between BFW and perceived personal shortcomings. The
findings are in line with the recent findings that indicate that people’s view on the free will
question can affect fundamental cognitive processes. Most importantly, belief in free will is
associated with an increased sense of agency (Lynn, Muhle-Karbe, Aarts, & Brass, 2014) and
Bias blind spot and free will beliefs                  25
self-efficacy (Baumeister & Brewer, 2012). In the similar vein, current findings support the view
that more people believe in free-will the less they perceive the personal shortcomings of the self
because of the agentic view that their own behavior is generated by themselves (e.g., desires,
goals), rather than by constraints. Across two studies, results confirm the hypothesis.
The exploratory hypothesis that tested for the relationship between free will beliefs and
magnitude of blind spot related to biases did not indicate conclusive support. However, we did
not find any effects to the opposite direction, but rather effects indistinguishable from zero. The
findings suggest that free will may not have a meaningful influence on the invidious distinctions
people make between their own and others’ susceptibility to bias. Previous work by Genschow et
al. (2017) finds that free will beliefs are positively correlated with correspondence bias. The
current finding suggests free will beliefs not have the same nature of the relationship with
individuals’ susceptibility towards other kinds of biases.
Findings of study 3 are in support of the pre-registered exploratory hypothesis that belief
in free was negatively associated with negative personality dimensions (snobbery, deceptiveness,
and selfishness). Findings are consistent with the theoretical view that belief in free will is
associated with moral responsibility. For example, Vohs and Schooler (2008) found that inducing
disbelief in free will increased participants’ cheating behavior. Similarly, Baumeister,
Masicampo, and DeWall (2009) found that an attenuated belief in free will reduce participants’
pro-social inclinations. Martin et al. (2017) find that free will beliefs positively related to harsher
punishments of unethical behavior. Negative personality dimensions included in the current study
do correspond to the moral responsibility in a person. However, we found no support for the
prediction that free will beliefs are positively correlated with positive personality dimensions.
Results suggest that belief in free will may not be associated with better than average effect in
Bias blind spot and free will beliefs                  26
regards to positive personality dimensions. The lack of support for this hypothesis is consistent
with the theoretical argument that free will underlies laypersons’ sense-making for accountability
and choice more so under negative circumstances (Feldman, Wong, & Baumeister, 2016).
However, the results of the correlation between free will beliefs and the extent of denial of bias is
positive and significant. 
In summary, results from the pre-registered extension hypotheses indicated that direction
of correlation holds in almost all cases: with a couple of exceptions (noted above). When the
exceptions occur to the pre-registered hypothesis, we did not find effects to the opposite
direction, but rather effects indistinguishable from zero. 
Conclusion
We aimed to replicate and extend previous findings of bias blind spot effect that refer to
the tendency to see bias in others while being blind to it in ourselves. For the most part, we
replicated the results reported by Pronin et al. (2002). The study contributes to the recent call for
systematic, large-scale, and preregistered replication and validation studies. Additionally, the
present investigation explored the relationship between free will beliefs and the tendency to
impute bias more to others than to the self is rooted. We extended the literature on bias blind spot
exploring the sources of bias blind spot (e.g., Pronin & Kugler, 2007). 

A Forensic Examination of China’s National Accounts 2008-2016: Growth overstated by more than 13pct

Chen, Wei, Xilu Chen, Chang-Tai Hseih and Zheng (Michael) Song. 2019. “A Forensic Examination of China’s National Accounts” BPEA Conference Draft, Spring. https://www.brookings.edu/bpea-articles/a-forensic-examination-of-chinas-national-accounts/

ABSTRACT: China’s national accounts are based on data collected by local governments. However, since local governments are rewarded for meeting growth and investment targets, they have an incentive to skew local statistics. China’s National Bureau of Statistics (NBS) adjusts the data provided by local governments to calculate GDP at the national level. The adjustments made by the NBS average 5% of GDP since the mid-2000s. On the production side, the discrepancy between local and aggregate GDP is entirely driven by the gap between local and national estimates of industrial output. On the expenditure side, the gap is in investment. Local statistics increasingly misrepresent the true numbers after 2008, but there was no corresponding change in the adjustment made by the NBS. We provide revised estimates of local and national GDP by re-estimating output of industrial, wholesale, and retail firms using data on value-added taxes. We also use several local economic indicators that are less likely to be manipulated by local governments to estimate local and aggregate GDP. The estimates also suggest that the adjustments by the NBS were insufficient after 2008. Relative to the official numbers, we estimate that GDP growth from 2008-2016 is 1.7 percentage points lower and the investment and savings rate in 2016 is 7 percentage points lower.

Greater male variability is currently universal in internationally comparable assessments; some of this heterogeneity can be attributed to some species universal mechanism or some other social/cultural phenomenon

Sex differences in variability across nations in reading, mathematics and science: a meta-analytic extension of Baye and Monseur (2016). Helen Gray, Andrew Lyth, Catherine McKenna, Susan Stothard, Peter Tymms and Lee Copping. Large-scale Assessments in EducationAn IEA-ETS Research Institute Journal 20197:2. https://doi.org/10.1186/s40536-019-0070-9

Abstract: A recent study by Baye and Monseur (Large Scale Assess Educ 4:1–16, 2016) using large, international educational data sets suggest that the “greater male variation hypothesis” is well supported. Males are often over-represented at the tails of the ability distribution despite similarity in measures of central tendency and the gradual closing of the attainment gap relative to females. In this study, we replicate and expand Baye and Monseur’s work, and explore greater male variability by country using meta-analysis and meta-regression. While we broadly confirm that variability is greater for males internationally, we find that there is significant heterogeneity between countries, and that much of this can be quantified using variables applicable across these assessments (such as test, year, male–female effect size, mean country score and Global Gender Gap Indicators). While it is still not possible to make any causal conclusions regarding why males are more varied than females in academic assessments, it is possible to show that some national level variables effect the magnitude of this variation. Results and suggestions for further work are discussed.

Introduction
Sex differences in cognitive abilities is a contentious issue, yet one that continues to draw the attention of the public and the research community alike. 21st Century society is motivated to ensure issues of equity between the sexes are adequately addressed, particularly within the sphere of educational opportunity (Marks 2008; UNESCO 2011). Despite best efforts however, inequities still exist internationally, for example, with females underrepresented in our most prestigious educational institutions and males overrepresented in school underperformance, particularly in core areas such as reading and mathematics (Baye and Monseur 2016; Dubet 2010; Jacobs 1996; Morgan and Kett 2003; Quinn and Wagner 2013).

While the evidence shows that the gap between men and women is closing on average across many educational outcomes (Hyde et al. 1990; OECD 2015) and, in some cases, it now favours women (Lietz 2006; OECD 2015), this shift does not appear to have translated directly into ensuring parity across higher professions and positions, a phenomenon which appears somewhat paradoxical. Baye and Monseur (2016) suggested that this may be due to the way in which sex differences have been historically examined, focussing on mean results which assume homogeneity of variance across the achievement distribution. In a study using international assessment data, they demonstrated that the magnitude of the sex differences in achievement across literacy, mathematics and science varied across the range of results, and that the largest differences are seen at the extreme tails of the distribution. Girls tended to outperform boys at both tails of the distribution on reading measures, and in the lower percentiles of mathematics and science, while boys outperformed girls in the higher percentiles of mathematics and science. While the differences at the top of the distribution were of note, they called attention to the fact that inequities in the lower percentiles of the distribution were much more striking.

Baye and Monseur (2016) also examined the variance ratios of boys and girls on these assessments and found that in 93% of cases, variances for boys were higher. The finding of greater male variances in assessments here is not in and of itself original and has been noted in studies for many decades (although rarely as a core focus). The “greater male variability” hypothesis in fact has its roots in the 19th century (Ellis 1894). However, if we are to understand differences between the sexes at different points of the distribution, we must attempt to determine how their respective distributions differ. It is to the issue of differences in variability, not average performance, that the rest of this paper attempts to address building on earlier work.

Male and female variability
Differences in the spread of scores between males and females have been noted in educational assessments for a long time, although often with contrasting findings. Maccoby and Jacklin (1974) showed that males were more variable than females in mathematical and spatial abilities, whereas variances showed parity in verbal measures. Feingold (1992) found larger male variances in the domains of general reasoning, mechanical reasoning, abstract reasoning, quantitative and spatial abilities, perceptual speed, memory and on verbal test batteries. Strand et al. (2006) found similar patterns in the domains of verbal, quantitative and non-verbal reasoning on a representative sample of 11-year olds in the UK, with greater male variances ranging between 7 and 17%. Similar results on U.S. students were found by Lohman and Lakin (2009) and later, Lakin (2013). IQ scores have also shown to reflect the same pattern (Johnson et al. 2008). Finally, assessments of non-cognitive and behavioural domains such as creativity (He et al. 2013; Karowski et al. 2016), sensation seeking (Cross et al. 2011), personality (Borkenau et al. 2013) and aggression (Archer and Mehdikhani 2003) appear subject to the effect. Combine these findings with the work reported earlier from Baye and Monseur and the fact that the above represents only a fraction of reported findings, one can see why many consider greater male variability to be ubiquitous.

Yet despite the volume of work related to differences in variances between the sexes, there has been little systematic attempt to explain this phenomenon (either partially or in its entirety). This is likely in part due to the contention that studies on sex differences in abilities tends to bring with it. Feingold (1992) noted that the explanation for greater male variability has become a polarised nature versus nurture debate. As a result, many empirical papers avoid proposing an explanation. Johnson et al. (2008) point out that although results have often seemed clear, studies are often attacked on methodological grounds pertaining to sample size, representativeness, sample selectivity and age amongst other things. While it is not our intent to repeat the full history of the greater male variability hypothesis (see Johnson et al. for an in-depth review) we will briefly consider some of the proposed explanations for this effect.

Explanations for greater male variability
As Feingold claimed, arguments regarding biological innateness are often invoked for theories of sex differences in cognitive and behavioural domains. Early theories (Ounsted and Taylor 1972) focused on the Y chromosome, claiming that differences in gene expression resulted in slower development and expressed more harmful as well as more beneficial traits, which would presumably lead to more variability in males. Gualtieri and Hicks (1985) suggested such differences could emerge from differences in the uterine environment, making males more differentially susceptible to physical and psychological disorders over the lifespan.

Evolutionary theories suggest that ancient adaptive mechanisms produced greater male variability to enhance survival in ancestral environments and that they are still in operation today. Evolutionary theories are based on sexual selection theory and parental investment theories (see Archer and Mehdikhani 2003 for a comprehensive review) and they would ultimately result in males showing greater variation across a range of traits in order to ensure reproductive fitness. Hill (2017) proposed two mathematical models simulating how one sex could have become more variable over evolutionary time if one sex in our ancestral past (presumably females in the case of homo sapiens, although Hill makes no explicit assumption) is more selective of the other for the purposes of mating, and that this greater variability will be independent of other measures of central tendency. Hill also suggested that in such circumstances where the selective sex is no longer being as selective, greater variability in the selected sex may in fact decline over successive generations. No direct test of this latter hypothesis has been made however.

While many support the biological and evolutionary basis for greater male variability, there are some shortfalls in this interpretation, as well as additional potential explanations as to why males are perhaps more variable. Miller (2001) claimed that susceptibility to defects resulting from prenatal conditions would only explain why males are overrepresented in the lower, not the higher tail of a distribution. As early as (1922), Hollingworth argued for an explanation based on gender roles, claiming that male employment, compared to the more restricted home role of women, allowed them the opportunity for greater diversification in education and environmental experiences. Noddings (1992) highlighted the issue of conformity, claiming that while most girls worked hard enough to avoid being in the bottom of the distribution in class, brighter girls are often pressured into not demonstrating the full extent of their abilities. Ceci et al. (2009) argued that biological accounts of differences in quantitative fields between the sexes are largely inconsistent and suggested that female preferences were a better explanation of underrepresentation in some professions. Critics of the evolutionary perspective also argue that if this phenomenon resulted from innate, evolved mechanisms, invariance of this effect across cultures would be expected. Several previous studies indicate that some nations show greater male variation, others greater female variation and many show homogeneity of variance (Feingold 1994). Feingold went on to attribute heterogeneity in his data to social and cultural factors rather than any innate biological mechanism. Feingold (1992) also argued that national test norms alone may not be sufficiently generalizable to afford definitive proof of a biological origin of greater male variability. However, more recent studies using international assessments such as PISA, PIRLS and TIMSS do seem to suggest that variability is greater for males in the domains of reading and mathematics across cultures (Baye and Monseur 2016; Machin and Pekkarinen 2008).

There has been some suggestion that elements of test design may also play a role in magnifying sex differences in terms of measures of central tendency and variances. Spelke (2005) claimed that supposed differences in ability, particularly in mathematics and science, resulted largely from item and test biases favouring males, and that research generally fails to support the greater male variability hypothesis in these domains. Lakin (2013) supports this to an extent, suggesting that changes to Cognitive Ability Tests (specifically, the introduction of new quantitative reasoning items with a lesser verbal load) may have been responsible for shifting more males into the upper echelons of the distribution compared with earlier versions of the assessment. Strand et al. however found few substantive sex differences related to item difficulties in non-verbal and verbal batteries and suggested that test construction was unlikely the root cause of differences in variability. They made a tentative suggestion that a speed-accuracy trade off favouring boys may account for some of the variability differences in quantitative domains, but cautiously note that that previous research has mirrored these effects in untimed assessments (such as Feingold 1992). Lakin also noted that the consistent trend of increasing variance ratios between cognitive ability tests at grades 4 and 7 is likely to be something more systematic than simple test design and potentially reflects changes to society in terms of educational opportunity and personal educational preferences. Arguments focussing purely on test construction and procedure are thus hard to substantiate in the current literature.

Machin and Pekkarinen (2008) highlighted a compositional effect of sex differences in central tendency and distribution of scores. In their analysis of TIMSS and PIRLS data in 15-year olds, they noted that greater male variance in maths was attributable to overrepresentation of males in the higher part of the test distribution, with males outperforming females on average. In reading, male overrepresentation was largely at the bottom of the distribution, with females outperforming males on average. Indeed, Nowell and Hedges (1998) found a correlation of 0.74 between variance ratios and male–female effect sizes. Baye and Monseur found a smaller overall correlation of 0.42. However, they noted that the strength of the relationship varied by the point in the distribution. At the 5th percentile, the relationship was 0.50. At the 95th percentile, this had declined to 0.31. These results seem to suggest that variability for males increases in line with superior female performance, particularly at the lower end of the distribution.

The current study
While Feingold’s work (1994) failed to show a consistent greater male variance in international test scores, this could be attributable to the methodology. He conducted a meta-analysis by searching the literature for reading, mathematics and spatial measures, which carries many issues with it including many different tests, test administrations, issues of representation etc. Baye and Monseur (2016), using more recently available international assessments (PISA, PIRLS and TIMSS) found different results, suggesting that greater male variability was effectively universal. They found that variances (on average) were 15% greater for males in reading, 12% greater in maths and 14% greater in science. Even using Feingold’s (1994) conservative estimate of any ratio falling between 0.90 and 1.10 as not representing evidence of greater variance, Baye and Monseur’s work is suggestive of greater male variability. The advantage of using these international assessments is that they are designed to be internationally comparable, with representative samples of children selected in each country and administered in a standardised fashion. This helps remove potentially confounding factors that may impact on assessment results.

However, Baye and Monseur’s work leaves many questions unanswered. How similar are countries to each other in terms of variance ratios, and are there some that are much more male biased than others? If countries vary in terms of male and female variances, are there any recorded factors that may account for this? Baye and Monseur did make some attempt to look at differences between primary and secondary school measures, as well as by IEA and OECD membership, but beyond this, no systematic heterogeneity analysis was conducted. Yet analysing heterogeneity is important and can be revealing. Furthermore, this international data could be linked to cross-country metrics that may elucidate meaningful patterns of variation. For example, Borkenau et al. (2013) showed that differences across countries in variances in personality were significantly linked to national measures of gender inequality and human development. Given earlier suggestions by Hollingworth (1922) that variances favouring males are largely due to gender roles, and later works (Ceci et al. 2009; Lakin 2013) suggesting that societal practices and female choice are likely to have a major impact on variance ratios, international indices of societal development, particularly forms of gender inequality, are potential sources that could be used to explain any cross-national heterogeneity. To our knowledge, this has not been examined in the context of large-scale international assessments.

In this study, we attempt to answer these questions and extend our knowledge surrounding the nature of greater male variability. We examined the same data sets used by Baye and Monseur, with the addition of more recent test administrations from years 2015 and 2016, to (1) replicate their findings using meta-analysis, (2) determine if greater male variability is homogenous both within and between countries and (3) quantify any meaningful sources of heterogeneity. For the purposes of the third aim, we link these data to international metrics on human progress (Human Development Index) and male–female participation in education, labour forces and politics (Global Gender Gap Index) as well as examining test specific factors such as grade, test, OECD membership, the size of the male–female difference at the mean and national means.

Method
Data sources
Data from three major international assessments were selected to allow an examination of variance ratios across countries: OECD PISA (Programme for International Student Assessment; 2000, 2003, 2006, 2009, 2012, 2015), IEA PIRLS (Progress in International Reading Literacy Study; 2001, 2006, 2011) and IEA TIMSS (Trends in International Mathematics and Science Study; 1995, 1998, 1999, 2003, 2007, 2008, 2011, 2015). These were selected due to having multiple testing points over time and having a wide coverage of countries across the globe. All data is freely available from the OECD website (http://www.pisa.oecd.org) and IEA Study Data Repository (http://rms.iea-dpc.org). Methodological information is available in the technical reports on each survey (Adams and Wu 2002; Martin et al. 2000, 2003, 2004, 2007, 2016; Martin and Kelly 1996, 1997; Martin and Mullis 1996, 2012; OECD 2005, 2009a, 2014, 2016; Olson et al. 2008).

International data on Human Development was also collected where available for each country. The Human Development Index (HDI) is made up of four sub-factors: expected years of schooling for children of school entry age, mean years of schooling for adults aged 25 and above, life expectancy and gross national income per capita (GNI). This data is freely available from the United Nations Development program website (http://hdr.undp.org/en/data).

International data on gender inequality was also gathered from the Global Gender Gap project. The Global Gender Gap Index (GGGI) is made of four sub-factors: economic participation, educational attainment, health and survival and political empowerment. Each factor represents an outcome and is measured on a scale of 0 to 1, where a score of 1 would represent parity between males and females. Data is freely available from the World Economic Forum’s website (http://reports.weforum.org).

Sample
Data from each country surveyed within each of the assessments was included in this analysis. For the purposes of this study, we used measures from three content areas: literacy, maths literacy and science literacy. In total, we included 564 cases for literacy, 1054 cases for mathematics literacy and 991 cases for science literacy gathered from over 100 nations worldwide (where each case represents a national test occurrence within a given year and within a specific content area). In terms of population size across all cases, in mathematics literacy it consists of 2,507,046 males and 2,512,273 females, for reading 1,471,698 males and 1,486,578 females and for science literacy 2,512,559 males and 2,515,645 females. It should be noted that for science literacy, we did not use data from TIMSS Advanced as these measures focussed on concepts from Physics only.

Data calculations
Statistics were calculated by generating means and standard deviations for males and females within each country for each measure within each assessment. These were calculated using each of the five plausible values within each database and aggregated according to the methodologies supplied by the OECD and IEA in their analyses manuals (OECD 2009b; Martin et al. 2016). Standard errors for these statistics were calculated using replicate weights within each database (80 Fay weights in PISA and 75 JK2 replicates in PIRLS and TIMSS). SPSS (V22; IBM Corp 2013) was used to calculate these statistics (see OECD 2009b; Martin et al. 2016 for technical details regarding the SPSS macros used to compute these statistics).

Variances were calculated from the standard deviations. The ratio of male to female variances was taken by dividing the male variance by the female variance. A variance ratio greater than one would indicate that the male variance is higher than the female variance. Variance ratios are a common method of examining variability between the sexes (see Hedges and Friedman 1993; Baye and Monseur 2016). In keeping with previous authors (Hedges and Friedman 1993; Katzman and Alliger 1992), but not Baye and Monseur (2016), ratios were logarithmically transformed to increase precision of the estimates and to avoid overestimation, as it ensures a normal distribution. Assuming that the log of the variances follows a normal distribution, the variances of these ratios were then calculated as:
v=2/(nf−1)+2(nm−1)
As we are examining variance ratios by country, some of the data points were combined for the purposes of the analysis. Countries such as Italy, Spain, Canada and the United States often report data for sub-regions but not consistently over assessments. These were collapsed for the purposes of this study. Where a nation has national and regional data within a given test administration, the subnational data points were used. China and the United Kingdom also report at the level of autonomous states (England, Scotland, Northern Ireland, Taipei, Machao, Shangai and Hong Kong). Countries falling into these states are denoted in the table but are not considered separately for aggregation. Assessments were considered together regardless of whether they were done in the primary or secondary years.

Meta-analysis
To examine the overall size of the variance ratio and to meaningfully quantify heterogeneity, meta-analyses were conducted using Comprehensive Meta-Analysis Version 3 (Borenstein et al. 2013). Many traditional analyses assume that effect size parameters are fixed and relatively homogenous. In this study, we are not assuming homogeneity of these parameters and are thus implementing a random effects model, assuming that effect size parameters are randomly sampled. The use of a random effects model is appropriate where heterogeneity is expected. In this study, we examined heterogeneity by country, whether the countries were OECD member states, test and grade.

Heterogeneity is examined by calculating Q statistics, which can be used to test for equality of effect sizes within and between analysis categories and follow the formulae below:
Q=∑i=1kw(di−d¯)2,
where w=1/v,v=(Nmale+Nfemale)/Ntotal+d2/2(Ntotal), and k is the number of effect sizes.
Q statistics follow a Chi square distribution of k − 1 degrees of freedom (Hedges and Olkin 1985). While significant Q statistics can detect the presence of homogeneity, they are not indicative of its magnitude. They are also sensitive to sample size (Hardy and Thompson 1998; Higgins and Thompson 2002) and its presence is generally expected when analysing large numbers of studies (Higgins 2008).

The mean of the log variance ratios, standard errors and confidence intervals for each country were then calculated (and presented in their un-transformed format for ease of understanding). For each country, we also tabulated the proportion of studies where; (1) the variances were significantly larger for males, (2) the variances were larger for males but not significantly so, (3) the variances were greater for females but not significantly so and finally (4) the variances were significantly greater for females.

Meta-regression
Meta-regression was used to explore and quantify potential sources of heterogeneity. We recorded the mean test score for each country in each year and calculated a weighted effect size of the gender difference between male and female means, as previous work has suggested that this effect size is related to the variance ratio (Baye and Monseur 2016). This was taken as the female mean subtracted from the male mean (a negative score therefore suggests higher scores for females). Using SPSS, this was converted into a standardised effect size (Hedges g) calculated from the effect size d multiplied by the correction factor J (correcting for small sample sizes):
d=μ1−μ2SDpooled
J=1−34df−1
Other additional moderators were derived from test administrations. Previous researchers (discussed earlier) have suggested that some differences may result from test design. As such, the test type, year, test grade and OECD membership were included as moderators to determine if these had a substantial impact on heterogeneity. Baye and Monseur (2016) found small differences in variance ratios between these variables and thus they may be contributing to some of the heterogeneity. Alongside these, the subfactors of the HDI and the GGGI were included to see if other country level contributing factors could account for variation across countries. As consistent data for both these indices is only available from 2006, meta-regression was performed only on cases from test administrations from 2006 onwards.

Results
Analysis of each content domain is presented separately. Countries with only one or two data points are included in the analysis although conclusions about the stability of their variance ratios must be treated cautiously. Variance ratios and their confidence intervals are presented in their un-transformed form for ease of interpretation. The percentage of cases that have a variance ratio below (significantly and non-significantly) and above (significantly and non-significantly) 1, with ratios above 1 representing greater male variance, are also presented. Q statistics and their significance are also reported for each nation.

Mathematics literacy
Table 1 shows the results for this analysis on international mathematics literacy data sources. Each of the 102 individual participating nations is listed in alphabetical order.
Table 1
Variance ratios, confidence intervals and heterogeneity statistics for countries in the domain of mathematics literacy

[tables]

These covariates predicted 31% of heterogeneity in Mathematics Literacy, 46% of the heterogeneity in Science Literacy and 54% of the heterogeneity in Reading. Many of the factors included in the model explain significant amounts of variance in effect sizes however, this varies by domain. By far the most significant predictor is the size of the gender difference in scores (across all three domains). As the gap becomes larger in favour of females, the variance for males increases. The mean score of the country is statistically significant for reading and science literacy but has a very small, positive impact. The same can be said for the test year in mathematics literacy. There are small and significant effects for the tests (with TIMSS and PIRLS showing slightly less male variance) but this is harder to interpret, as it is confounded by age. HDI indicators seem to have little impact on variance ratios, although GNI has a very small positive but statistically significant effect on mathematics literacy and science literacy. GGGI indicators have a stronger, negative impact on national variance ratios however. Countries with higher Economic Participation for women have ratios favouring females across all domains. Better Educational Attainment for women significantly increases the ratios in favour of males however in mathematics literacy and science literacy. Increased political empowerment for women also seems to increase variances for females in literacy.

Discussion
Results broadly confirm the previous works of Baye and Monseur (2016) and suggest that male variances are greater than female variances internationally. This was largely expected as, although the methodology differed, most of the data used in this study was the same. Baye and Monseur showed variances for males were greater by 15% in reading, 12% in maths and 14% in science. Our results indicated that these ratios are 16%, 12% and 13% respectively, and suggest that the inclusion of more recent international surveys has not altered them substantively. Similarly, the correlation between male–female effect sizes and variance ratios was in line with those found by previous authors, with superior female performance increasing the gap in variance between the sexes. As such, we can broadly support the findings of past research and conclude that over the studied period, male variances in the domains of reading, mathematics literacy and science literacy are almost universally greater.

However, these results suggest that we can take this conclusion a step further. Feingold (1994) suggested that a difference of about 10% in variance ratios should be considered a substantive difference. Tables 1, 2 and 3 clearly show that for most countries engaging with PISA, PIRLS and TIMSS assessments, male variances are greater by often more than this threshold in all three domains. There are no geographical areas in this study that show significantly greater female variances. It would seem therefore that the question currently should no longer be, do male and female variances differ, but by how much more varied are males compared with females?

While in over 95% of cases, males show greater amounts of variance, there is a significant heterogeneity in these results, both within and between countries. While we can say with confidence that males are certainly more varied and generate a fairly precise estimate of a global average, we cannot come to an absolute value for each country individually and must contend with a large amount of dispersion. This dispersion is telling however and shows that not only do countries differ (significantly in some cases, as is evident in Figs. 1, 2 and 3) but that they vary internally as well. There is a significant amount of heterogeneity across these data in most countries examined in this study which requires explaining.

Our meta-regression within each domain has gone some way in explaining close to half of the heterogeneity observed in the dataset for reading and science literacy and about a third for mathematics literacy. Some of the findings are harder to interpret than others. The variable with the largest impact is the male–female effect size. This is the most substantive factor across all three domains and suggests that as girls outperform boys, the variability of boys increases. This seems to support earlier works that demonstrated a correlation between effect sizes and variance ratios (Baye and Monseur 2016; Nowell and Hedges 1998). The mean score for the country also has a significant albeit smaller impact in the same direction for science literacy and reading. Countries that perform better on average are therefore more likely to have greater variability for boys.

PISA tests appear to result in slightly more variance for males than TIMSS and PIRLS. Baye and Monseur (2016) found slightly smaller ratios in the primary years across all three domains. As TIMSS and PIRLS assess younger children, it may be that this simply reflects an age or maturity effect. However, we cannot rule out that the actual tests themselves are not causing some of the heterogeneity or, that there may be a compositional effect between the two.

Interestingly, most of the HDI indicators were not significantly predictive of variance ratios across domains. The exception to this appears to be the GNI indicator (an adjusted form of GDP per capita) for mathematics literacy and science literacy but not reading. Reading is a specific skill that requires mastery and is often contingent on home environments for reinforcement. While this is to an extent true of basic mathematical concepts, later mathematics and science are likely tied more strongly to whatever specific curriculum is delivered, and this is largely coordinated at a national level. This may explain why national wealth may impact more upon maths and science as opposed to reading. However, it should be noted that, despite its statistical significance, it has only a minute impact on increasing male variance.

Measures from the Global Gender Gap Index however seem to have a larger impact on variance ratios. Increasing female economic participation appears to increase levels of female variance across all three domains. This suggests that countries actively incorporating more women into the labour force has an impact on educational outputs. Increased political empowerment for women also increases female variances in reading. Increased educational attainment for women has mixed impacts however. It has a significant effect of increasing male variances in mathematics and science but a non-significant effect of increasing variance for women in reading. Taken together, it suggests that cultural practices tied to increasing female participation generally appear to increase variances for females and suggests that greater male variance in educational outcomes may be practically reduced on national levels. While this study cannot isolate what specific national level practices are responsible for this, it does lead to interesting further questions regarding the processes underlying male/female variability.

The year of the test also had a very small but statistically significant effect on variance ratios in mathematics literacy. As with the test variable itself, why precisely this should be the case is difficult to rationalise. As mentioned earlier, there could be specific test administrations which have differences that create a small, positive effect. Alternatively, it could be that national educational systems have been adapting educational practices in order to improve their position in international rankings, and that these new practices are impacting upon the spread of scores. From this data alone, we can only speculate on the specifics as to why this may be the case.

Limitations and future work
There are several limitations to the data and the procedure we have used to explore it. First, a meta-analysis of international assessments such as PISA, PIRLS and TIMSS, while it controls for many extraneous variables not possible to account for in a meta-analysis via a literature search, does limit generalizability to alternative educational assessments. There could be something specific to these assessments that creates this effect. A limitation perhaps related to this applies to the assessments themselves. In PISA, the content being assessed is heavily based in literacy abilities. Even mathematics and science components are rooted in the ability to read and poor readers are unlikely to achieve if they cannot interpret the questions posed. As is evident from Table 2, the domain with the greatest amount of male variability is reading. As such, it is possible that mathematics and science show comparable overall ratios simply because they are rooted heavily in the ability to read. It is interesting to note that previous works using different assessments have shown greater variabilities in quantitative domains compared to verbal ones (Lakin 2013; Lohman and Lakin 2009). Thus, what this data may perhaps be showing is the greater variability in reading generally. This is still important and would pose the question ‘why are males more variable at reading’ but we must therefore be cautious regarding the conclusions we draw from the mathematics and science domains.

This study tentatively suggests (as does Baye and Monseur 2016) that age may be a factor, and that variability for males increases as candidates get older. To our knowledge, no study specifically examines this, either longitudinally or cross-sectionally (with perhaps the exception of Lakin 2013). Alternatively, attempting to quantify nation specific factors that could be included in additional regression analyses may be a future avenue worth exploring (particularly considering the impact of GGGI variables on ratios), potentially allowing us to quantify greater levels of heterogeneity in these results.

A final avenue of exploration would be to examine this effect over additional academic assessments. Research historically focuses on core domains of reasoning (Baye and Monseur 2016; Lohman and Lakin 2009; Strand et al. 2006). While this is important, do we get similar patterns across curricular subject examinations (anything from art to zoology), or different modes of assessment (pencil and paper tests compared to practical performance assessments)? These are often studied less, in part due to reasons of sample representation, or the fact that specific subjects are often self-selecting. As it stands from the data and the literature reviewed here, we would expect to see similar patterns across assessments generally. It would be telling if this was not the case. If there are exceptions, what are they and why do they differ?

Implications for theory and policy
From a theoretical perspective, we cannot contribute causal explanations for why males are more variable. Data suggests the effect is almost universal, which, while supportive of biological and evolutionary theories, doesn’t rule out specific cultural, educational, political, social or religious practices. Indeed, the fact that we can quantify substantial variation as dependent on increased female participation in society suggests that, at least in educational outcomes, it is not necessarily the case that males should vary more.

However, without a clear understanding of why males vary more and how this difference is maintained, we acknowledge that a meaningful discussion regarding what can be done to ensure parity is difficult. Increased female participation in the economy, education and political empowerment significantly reduce the size of the discrepancy in variances between males and females across the three educational domains studied here. If these increase, we might expect the variance gap to decrease. Which specific practices within countries are enabling this however are not discernible from the existing data, and more comparative, in-depth work within nations (with closer attention to specific educational practices) would be required before specific policy recommendations could be formulated to ensure parity between males and females across the ability distribution.

Differences in the spread of abilities are important for society. If, for example, we want to increase the representation of women in top positions and educational institutions, so that parity between the sexes exists at this level, it is important that males and females are equally represented in the higher percentiles of whatever qualifications or ability metrics that constitute the selection processes. Similarly, the large gap in reading ability between boys and girls in the lower percentiles (Baye and Monseur 2016) suggests that some boys are likely to be at a serious disadvantage in later education (and potentially later life outcomes). Whilst implementing measures that strive for parity in the right tail of the distribution are important, we must also be mindful to not neglect the left.

Conclusions
Our analysis seems to suggest that greater male variability is currently universal in internationally comparable assessments implemented over the past decade. However, this effect is far from homogenous, and there are quantifiable differences that exist over nations. Furthermore, some of this heterogeneity can be attributed to some yet unspecified practices or policies targeted at increasing male–female equality, general male–female performance as well as potentially the age of candidates and the type of test. Further work however is required to examine these factors in more detail, and analyses within nations may be informative to examine more specific practices that can explain national patterns. Comparative work examining high and low scoring GGGI countries may be informative in this endeavour. In doing so, it may be possible to determine if the root cause of these differences in distributions are attributable to some species universal mechanism or some other social or cultural phenomenon.

Exposure to female estrous, a natural rewarding experience, alleviates anxiety & depression, & is beneficial to recovery following transient ischemic stroke in male mice

Exposure to female estrous is beneficial for male mice against transient ischemic stroke. Yuan Qiao, Qing Ma, Haifeng Zhai, Ya Li & Minke Tang. Neurological Research, Feb 27 2019, https://doi.org/10.1080/01616412.2019.1580461

Objective: Exposure to female estrous, a natural rewarding experience, alleviates anxiety and depression, and the contribution of this behavior to stroke outcome is unknown. The aim of this study was to evaluate whether exposure to female estrous is beneficial to recovery following transient ischemic stroke in male mice.

Methods: Cerebral ischemia was induced in male ICR mice with thread occlusion of the middle cerebral artery (MCAO) for 30 min followed by reperfusion. MCAO mice were randomly divided into MCAO group and Estrous Female Exposure (EFE) group. The mice in the EFE group were subjected to estrous female mouse interaction from day 1 until the end of the experiment. Mortality was recorded during the investigation. Behavioral functions were assessed by a beam-walking test and corner test from day 1 to day 10 after MCAO. Serum testosterone levels were analyzed with ELISA, and the expression levels of growth-associated protein-43 (GAP-43) and synaptophysin in the cortex of the ischemic hemisphere were determined by western blot on day 7 after MCAO.

Results: Exposure to female estrous reduced the mortality induced by cerebral ischemic lesions. The beam-walking test demonstrated that exposure to female estrous significantly improved motor function recovery. The serum testosterone levels and ischemic cortex GAP-43 expression were significantly higher in MCAO male mice exposed to female estrous.

Conclusion: Exposure to female estrous reduces mortality and improves functional recovery in MCAO male mice. The study provides the first evidence to support the importance of female interaction to male stroke rehabilitation.

Abbreviations: GAP-43: growth-associated protein-43; SYP: Synaptophysin; MCAO: middle cerebral artery occlusion; OVXs: ovariectomies; CCA: common carotid artery; ECA: external carotid artery; EFE: estrous female exposure; TTC: 2,3,5-triphenyltetrazolium chloride; PAGE: polyacrylamide gel electrophoresis; PVDF: polyvinylidene difluoride; ANOVA: analysis of variance; LSD: least significant difference

KEYWORDS: Exposure to female estrous, transient ischemic stroke, GAP-43, SYP

Females having a younger brother: Adult earnings are about 7 pct less in adulthood; parents have lower academic expectations for the girls, & the girls develop more traditionally feminine roles

The Brother Earnings Penalty. Angela Cools, Eleonora Patacchini. Labour Economics, https://doi.org/10.1016/j.labeco.2019.02.009

Highlights
•    We examine the impact of having a younger brother on females’ adolescent environment and adult earnings
•    We use unique longitudinal data on a recent cohort of U.S. women
•    Girls with a younger brother earn about 7 percent less in their adulthood
•    Parents have lower academic expectations for girls with a younger brother
•    Girls with a younger brother develop more traditional gender roles and behaviors

Abstract: This paper examines the impact of sibling gender on adolescent experiences and adult labor market outcomes for a recent cohort of U.S. women. We document an earnings penalty from the presence of a younger brother (relative to a younger sister), finding that a next-youngest brother reduces adult earnings by about 7 percent. Using rich data on parent-child interactions, parents’ expectations, disruptive behaviors, and adult outcomes, we provide a first step at examining the mechanisms behind this result. We find that brothers reduce parents’ expectations and school monitoring of female children while also increasing females’ propensity to engage in more traditionally feminine tasks. These factors help explain a portion of the labor market penalty from brothers.

Null Effects of Game Violence, Game Difficulty, and 2D:4D digit ratio, thought to index prenatal testosterone exposure, on Aggressive Behavior

Null Effects of Game Violence, Game Difficulty, and 2D:4D Digit Ratio on Aggressive Behavior. Joseph Hilgard et al. Psychological Science, March 7, 2019. https://doi.org/10.1177/0956797619829688

Abstract: Researchers have suggested that acute exposure to violent video games is a cause of aggressive behavior. We tested this hypothesis by using violent and nonviolent games that were closely matched, collecting a large sample, and using a single outcome. We randomly assigned 275 male undergraduates to play a first-person-shooter game modified to be either violent or less violent and hard or easy. After completing the game-play session, participants were provoked by a confederate and given an opportunity to behave aggressively. Neither game violence nor game difficulty predicted aggressive behavior. Incidentally, we found that 2D:4D digit ratio, thought to index prenatal testosterone exposure, did not predict aggressive behavior. Results do not support acute violent-game exposure and low 2D:4D ratio as causes of aggressive behavior.

Keywords: violent video games, aggressive behavior, digit ratio, Bayesian analysis, open data, open materials