Wages and CPI – not what you think

Wage growth data is a focal point for stock markets at the moment. Does it matter?

One of the “reasons” given by the financial media for the recent sell-off in global stock markets was a larger-than-expected increase in U.S average hourly earnings. Average hourly earnings increased by 2.9% on an annualized basis in January, versus 2.6% expected. The stock market, which had already peaked out six days before the numbers were released, nevertheless increased its pace of decline, prompting media and economists to link the two together. The belief is that people sold stocks because an increase in average earnings will make it more likely that consumer prices will rise, and therefore that the Federal Reserve will increase interest rates more than previously thought. This sounds logical and rational. But passing over the fact that people sell and buy stocks for thousands of different “reasons” each day; does it have any merit?

Underlying the concern over higher wage growth is the belief in a link between wage growth and consumer prices. This, however, is a distortion of an economic “rule” that many people think of when focusing on wage growth. William “Bill” Phillips was a New Zealand-born economist who spent most of his career at the London School of Economics. He is best known for his work on the link between levels of employment in an economy and wage growth. This “Phillips Curve” states that as levels of unemployment drop, wage growth increases. In the U.S., there is some evidence to suggest that this is true. The rolling 36-month correlation between the U.S. unemployment rate and the annualized growth rate of average hourly earnings fluctuated between -0.65 and -0.87 between 1994 and 2008. A negative correlation like that suggests that the Phillips Curve has some merit. Prior to 1994 though, it was as high as +0.65 in the 1970s and +0.42 in 1990. A positive correlation between unemployment and wage growth suggests that the Phillips Curve has no merit. The U.S unemployment rate topped out in 2010, and since 2014, has continued to fall whilst wage growth has been relatively subdued. This has caused the rolling 36-month correlation to move from -0.70 to -0.13, and led many to claim that the Phillips Curve is, once again, broken.

All of which would be fine, but Bill Phillips did not mention anything in his work about the link between unemployment and what people refer to as inflation – the annualized growth rate of the consumer price index (CPI). That notion was made popular by the famous monetary economist, Milton Friedman, in the late 1960s. Friedman’s idea was that as unemployment fell, CPI would rise, and vice versa. But here again, the relationship as measured by cold, hard, correlation data has swung between positive and negative over the decades, having been as high as +0.63 in 2000 (both unemployment and CPI coming down) and as low as -0.76 in 2009 (unemployment going up and CPI coming down).

Thus, the historical evidence suggests that the relationship between unemployment and wage growth is patchy, as is the relationship between unemployment and CPI. So what of the alleged relationship between wage growth and CPI that appears to have spooked the equity markets?

The chart below shows the annualized rate-of-growths of average hourly earnings and the CPI. The lower panel shows the rolling 36-month correlation between the two series. If higher wage growth is supposed to lead to higher CPI growth, we would expect to see a positive correlation. However, we can see that, since 1968, the correlation has swung between a low of -0.79 (made in 1972) and a high of +0.98 (made in 1983). The correlation has been in positive periods eleven times, and in negative periods ten times. And, in fact, the average correlation between wage growth and CPI since 1968 turns out to be +0.24. That is, on average, hardly any relationship at all.

The conclusion we must reach is that, whilst it may sound plausible that there is a relationship between unemployment, wage growth and consumer prices, an examination of the actual data tells a different story. There is no guarantee that higher wage growth will feed into higher CPI.

Most importantly, what this further underscores, is that most prognosticators are operating under wrong assumptions when it comes to what correlates with CPI. As such, their assessments regarding the outlook for consumer prices appear even less relevant.

US average hourly earnings YoY%