The growth in income disparity in the United States has become a big story. Many analysts point to the mid 1970’s as the beginning of the current trend as, since then, median real wages have stagnated and most of the benefits of productivity growth have accrued to the top of the income scale. 1975 was also the year that the Social Security Administration began publishing its annual Cost of Living Adjustments (COLAs) governing benefits increases for retired seniors, and the mid 1970’s were a time of major work on the government’s Consumer Price Indices as they evolved into the forms we are familiar with today.

Is this just coincidence or is there a causal relationship between the introduction and media penetration of these regularly published Cost of Living Adjustments and CPIs and flattening real median income growth?  We believe the there is a causal relationship. And in this brief article we will suggest that:

1.  The slow-down in United States production wage[1] growth coincided with Social Security’s introduction of its annually published Cost of Living Adjustments (COLAs)[2].

2.  Beyond just slowing down, beginning at this time production wage growth began to more closely track CPI/COLA[3] index growth, whereas in prior periods it had more closely tracked growth in GDP per Capita.

3.  A material contributor to this slowdown and change in wage growth patterns may have been the introduction of COLA itself, which provided a new, aptly titled, and singularly available anchor point for corporations and employees to use in setting annual wage growth targets, and demonstrated the broad utility of the CPI in adjusting cash compensation. We should not underestimate the power of the COLA reference point in impacting wage negotiations.

4. A simple, effective way to combat the downward wage growth bias contributed to by COLA is to publicize a new index, which we call the Staying Even Index[4] (SEI). The SEI tracks growth in U.S. GDP per Capita, and is designed to help companies and individuals know how much wage growth is necessary to keep up in the expanding U.S. economy, and to re-link wage growth expectations with growth in GDP per Capita.

1.  The Slow-Down in the Growth of Production Wages Coincided with the Social Security Administration’s Introduction of its Annually Published Cost of Living Adjustments (COLAs)[5] and refinements/deepening CPI metrics.

In 1972 Congress passed legislation mandating that, beginning in 1975, the Social Security Administration raise Social Security benefits through an annual Cost of Living Adjustment tied to the Consumer Price Index. Thus was born COLA and its mid-1970s birth year marks the period when median wage growth began to markedly trail growth in GDP per Capita (Figure 1), which it had closely tracked from at least 1950.

During this same period, the Bureau of Labor Statistics revamped its approach to collecting data for the Consumer Price Index (1972-1974 consumer expenditure survey) and in 1978 published its new CPI-U (all urban consumers), which represented a set of major changes in data collection and involved the Census Bureau for the first time.

In short, the mid-1970s saw the introduction of a major Cost of Living Index based upon the CPI that has since been published annually and saw major re-work of inflation metrics published by the government. And, median wage growth began to slow.

Figure 1: 1950-2014 Trends in GDP per Capita and Production Wages

2. At the Same Time, Production Wage Growth Patterns Changed from a Pattern of Growing with GDP per Capita to a Pattern of Growing with COLA/CPI.

To better understand and zoom in on the dynamics of median wage growth, we have divided the postwar period into two. This first is from 1950-1974 (Figure 2), and the second from 1975 to 2013 (Figure 3). We have indexed GDP per Capita, production wage, and CPI/COLA growth to 100 at the beginning of each period and a striking pattern emerges:

• From 1950 to 1974, production wages grew large in line with GDP per Capita and well above CPI/COLA growth (Figure 2).

• This pattern changed significantly around 1975. From 1975 to 2013, production wages grew substantially in line with CPI/COLA, well below GDP per capita growth (see Figure 3).

• What we see between Figures 2 and 3 then, is a striking change in the pattern/dynamics of wage growth, from one in which wages grew largely with GDP per Capita, to a new pattern in which wage growth patterns change and median wages began growing substantially in line with CPI/COLA, and well below GDP per Capita, and this happened right after COLAs were introduced in 1975.

  Figure 2: From 1950 to 1974, Production Wages Grew in Line with GDP per Capita

          Figure 3: From 1975-2013, Production Wages Grew in Line with CPI/COLA, Well Below GDP per Capita Growth

3.  The Introduction of COLA Itself May Have Been A Material Contributing Factor to this Wage Growth Pattern Change

While some may quibble with details of the data, our conclusions until this point should be relatively uncontroversial. It is this next step in our logic that may be. Many factors have been cited to explain the change in wage growth patterns beginning in the mid-1970s, from the decline of unionization to increasing benefits[6] costs to global wage competition. We believe a material contributor to this change may have been the introduction of COLA and the deepening of the CPI itself – which created strong reference points for wage negotiations and also some confusion about how much wage growth was necessary to keep up in the growing U.S. economy.  But first, let's recap what we have shown so far:

• The Social Security Administration began publishing annual COLA figures in 1975, and at the same time the Bureau of Labor Statistics and the Census Bureau reshaped the Consumer Price Index into largely the form we use today.
• In the 2 ½ decades until 1975, median wage grew largely in line with growth in GDP per Capita.
• In or around 1975, median wage growth trends materially changed and wage growth largely tracked COLAs/CPI for the proceeding 30+ years.

Is it then reasonable to suggest that the introduction of annually published COLA data and the accompanying increase in focus on CPI in itself contributed materially to this change in income growth patterns, and, in consequence, to the increasing wage disparities in the U.S. economy? While the correlations are powerful and make their own case, what might have happened?  We have a few thoughts:

• • The emergent COLA index began to serve as a mutual and “easily understood” reference point for individuals and corporations. Indeed, for a job that may change little from year to year (aside from an average productivity increase of a couple of points), it would seem “fair” to award a wage increase equivalent to the increase in “living expenses” born by the employee. And, the systematic application of CPI/COLA to social security payments would have legitimized its use in these contexts.

  • The source of the index – the government – and the title of the index – “Cost of Living Adjustment” – both suggested trustworthiness and suitability.

  • The increasing use of CPI/COLA to govern contractual price increases with customers reinforced the sense it would also be appropriate to use with employees. Indeed even prior to the implementation of Social Security COLA, an estimated 4.4M union employees were under collective bargaining agreements that linked wage growth to CPI.

  • Media readily embraced the index and no other comparable index existed that would suggest cost of living adjustments would not be sufficient for workers to “maintain their economic place” in society – i.e, to stay even with those around them.  

  • Employees misunderstood the indices to mean that wage growth equivalent to COLA/CPI would be sufficient for them to “keep up”.  This is just not true, as COLA/CPI growth significantly lags growth in GDP per Capita, meaning those who receive COLA-based raises fall a little bit behind each year. Over the course of decades, this impact is very large - a person earning GDP per Capita-based raises since 1975 would be earning 56% more today than a person who received COLA-based raises. Indeed, while many are concerned about income inequality, the fact is that no simple, widespread source of data tells individuals how much income they need to keep up (defined as enough income to avoid growth in income inequality). 

4.  A simple, effective way to combat the downward wage bias caused by the annual publication of COLA is to create a new index which we call the Staying Even Index (SEI), which re-links wage growth expectations to GDP per Capita Growth.

As we’ve discussed, we believe that CPI/COLA has acted as a readily accessible “anchor point” for annual wage discussions and decisions and has resulted in slower wage growth than would have otherwise occurred. Accordingly, the simplest way to combat its effect in this arena is to introduce a new, media-friendly index, which we call the Staying Even Index (SEI)[7] and which tracks growth in GDP per Capita.

Individuals who increase their income in line or greater than the SEI will be growing their adjusted share of the U.S. economy, and those who fall short of the SEI will be falling behind. Awareness and availability of this index, in and of itself, could help reverse very troubling multi-decadal trends in income distribution by creating new wage growth expectations among both employers and employees. 

We welcome your feedback on our perspectives. Please send comments to and feedback to . Please also visit our website at stayingeven.com, and follow @stayingeven on twitter.