Additional assets 40419


Additional assets 40421

– October 6, 2015

The Social Security Administration determines the COLA based on the Consumer Price Index for Urban Wage Earners and Clerical Workers (CPI-W) compiled by the Bureau of Labor Statistics. This index is different from the Consumer Price Index for all Urban Consumers (CPI-U), which is the widely reported measure of inflation. Most years, the two indexes track each other closely, but not this year. In the 12 months to August 2015, the CPI-W fell 0.3 percent, but the CPI-U increased 0.2 percent. The falling price of motor fuel had a larger impact on the CPI-W than on the CPI-U, causing the disparity. With small overall changes in the indexes this year, this difference matters.

We expect that once the final data are released, the change in the CPI-W will be negative and the COLA will be zero. When prices drop, the law stipulates that the COLA should be zero. This happened in 2010 and 2011 (see chart).

Does COLA compensate retirees adequately?

The intent of the automatic COLA is to compensate Social Security recipients when the purchasing power of their benefits falls due to rising prices. If prices fall, on the other hand, the purchasing power of benefits increases, making a zero COLA more than sufficient. But is CPI-W, the index that underlies the COLA, the right measure of the prices faced by retirees?

The CPI-W reflects the prices of goods and services purchased by a typical household of urban wage earners. This demographic covers about 28 percent of total U.S. population. In the past 12 months, these prices have dropped, driven mainly by falling energy costs. Over this period, gasoline prices fell more than 20 percent and energy costs overall fell 15 percent.

Retirees have different spending patterns than urban wage earners. For example, they consume more medical care but spend less on education and gasoline than urban wage earners. In an ideal world, the Social Security COLA would be based on an index that reflects the spending pattern of retirees. Such an index does not exist.

The closest we can come is an experimental Consumer Price Index for Americans 62 year of Age and Older (CPI-E), which the BLS has compiled since the early 1980s. This demographic group does not match retirees exactly, since many people in their 60s still work. In the 12 months to August, the CPI-E increased 0.65 percent. If we take this as an approximation of the price increase faced by retirees, a zero COLA would not fully compensate them. In the past, the COLA fell short of the change in the CPI-E in some years and exceeded it in others.

There is yet another way to look at prices that are relevant for Social Security recipients. Imagine a person who relies on Social Security benefits to cover only daily expenses – food, utilities, phone bill, cleaning supplies, gasoline, prescription drugs, and the like. Other major expenses (such as housing, major purchases, a car, etc.) are taken care of from some other source. The house might be paid off, eliminating monthly mortgage payments. For such a person, the purchasing power of Social Security benefits depends on how much the prices of everyday goods and services changed over the year. AIER has a measure that reflects this – the Everyday Price Index (EPI).

Over the 12 months to August, the EPI fell 2.8 percent, driven mainly by the fall in fuel prices. Using this measure, a zero COLA would more than compensate people who use Social Security benefits to cover the costs of everyday products.

Consequences of zero COLA beyond Social Security benefits

If the COLA is zero, the law provides that the maximum earnings subject to Social Security tax stay constant. In 2015, the maximum amount of taxable earnings is $118, 500, and this will remain the cap for 2016 if the COLA is zero. Other automatic adjustments linked to the COLA include Supplemental Security Income benefits, retirement earnings exempt amounts, and the earnings required for one Social Security credit. A zero COLA will mean that all these will stay fixed in 2016.

A zero COLA also has implications for the premiums that retirees pay for Medicare Part B, which covers physicians’ services, preventative care, lab tests, and the like. This connection is hidden within the fine print in the law and is called the hold – harmless provision. It provides that after paying any increase in Medicare premiums, no individual’s Social Security benefits can fall, year to year. (For most Medicare recipients, Medicare premiums are deducted directly from their monthly Society Security benefit check.)

In practical terms, Medicare premiums of those covered by the hold-harmless provision can increase no more than the dollar amount of the COLA increase of the Social Security benefit. If the COLA is zero, Medicare premiums stay fixed. The hold-harmless provision does not apply to premiums for Medicare Part D or Medicare Advantage, which can increase, although the Centers for Medicare & Medicaid Services (CMS) currently projects they will remain the same in 2016.

The hold-harmless provision covers the majority of Medicare beneficiaries but not everyone. It does not cover new enrollees in Medicare in 2016, high-income Medicare beneficiaries, and low-income individuals who are eligible for both Medicare and Medicaid, who have their premiums paid by Medicaid. It also does not cover those Medicare beneficiaries who are not receiving Social Security benefits. All in all, about 30 percent of Medicare beneficiaries are not covered by the hold-harmless provision.

No provision of the law can prevent the costs (as distinct from premiums) of the Medicare program from changing. Medicare Part B premiums are calculated every year to cover a quarter of the expected costs of the program. (The other three quarters is covered from the general tax revenues.) The premiums have stayed constant, at $104.90 per month, since 2013, but they rose every year before then.

If the total estimated costs of Medicare Part B for 2016 rise, the total collected premiums will have to rise as well. Since the premiums for those beneficiaries protected by the hold-harmless provision cannot increase, the premiums for those not protected by it will increase much more than they would have otherwise.

New enrollees, who usually constitute just a few percent of all Medicare beneficiaries, will see the largest increases in Medicare premiums, if there are any. In 2010, when hold-harmless provision kept Medicare premiums constant for most, new enrollees faced premiums about 15 percent higher than the existing beneficiaries. Back then it led to an uncomfortable situation of people with identical incomes paying very different premiums for Medicare depending on the date of their first enrollment (see Table 1).

High-income beneficiaries (defined as those making more than $85,000 per person or $170 000 per couple) account for about 5 percent of all beneficiaries. Their income-based Medicare premiums are already much higher than standard premiums, and they would increase more under a zero COLA.

Low-income individuals, who account for about 20 percent of Medicare beneficiaries, have their premiums paid by Medicaid. Since this group is the largest among those not protected by the hold-harmless provision, most of the increase in the premiums in 2016, if there is one, will be paid for by the taxpayers who finance Medicaid. This is in addition to the 75 percent of all Part B costs normally covered by the taxpayers.

Medicare Part B premiums for 2016 have not been determined yet. CMS makes the official determination, which is usually published in the Federal Register in early November and disclosed on the web at

Polina Vlasenko, PhD

Get notified of new articles from Polina Vlasenko, PhD and AIER.
AIER - American Institute for Economic Research

250 Division Street | PO Box 1000
Great Barrington, MA 01230-1000

Contact AIER
Telephone: 1-888-528-1216 | Fax: 1-413-528-0103

Press and other media outlets contact
[email protected]

Editorial Policy

This work is licensed under a 
Creative Commons Attribution 4.0 International License,
except where copyright is otherwise reserved.

© 2021 American Institute for Economic Research
Privacy Policy

AIER is a 501(c)(3) Nonprofit
registered in the US under EIN: 04-2121305