By Frances Coppola
The trade-off works like this: When unemployment is low, employers have to offer higher wages to attract workers from other employers. This increases their costs and hence forces them to raise prices. Thus, low unemployment causes higher inflation. Rising inflation tends to mean a falling exchange rate, because the purchasing power of the currency is being eroded. Thus, if low unemployment feeds through into rising inflation, the dollar’s currency exchange rate tends to fall.
The “natural” or “neutral” rate of unemployment, u-star, also known as the “non-accelerating inflation rate of unemployment” (NAIRU), is the unemployment rate at which inflation is stable and the economy is running at full potential. It is one of the “three stars” that govern Fed monetary policy decisions and hence influence the dollar’s exchange rate, the others being the “neutral” rate of inflation, pi-star, and the “neutral” interest rate, r-star.
Historically, NAIRU has been between 4 and 6 percent.1 But unemployment is now well below that, and yet inflation is still barely reaching the Fed’s 2 percent target.2 Many economists now think that the advance of technology and the changing nature of work mean that NAIRU is much lower than it was in the past. Some go even further, saying that in today’s low-interest-rate world, the relationship between unemployment and inflation no longer holds.3 As unemployment is a key input to Fed monetary policy decisions, the new uncertainty about its level and its role in the economy makes the path of interest rates, and hence the likely level of the dollar exchange rate, difficult to predict.
The Phillips Curve, named after its inventor, the British economist A.W. Phillips, plots the relationship between unemployment and inflation. Usually, this chart shows a distinctly negative correlation, meaning that as unemployment rises, inflation slows. However, the stagflation of the 1970s showed that inflation could rise even when unemployment was high: in 1975-6, for example, the British pound’s exchange rate collapsed4 due to double-digit inflation5 despite unemployment of 5.7 percent, at that time the highest since the 1930s.6 So the economists Milton Friedman and Edmund Phelps, working independently, both amended the Phillips Curve to take account of workers’ expectations of both rising wages and rising prices – and in so doing, introduced the concept of a “natural” or “neutral” rate of unemployment.
Milton and Phelps said that the trade-off between unemployment and inflation would only hold for short periods. If the government introduced expansionary measures to reduce unemployment below its natural level, rising wages would encourage people to work more, so unemployment would fall in the short term. But rising demand for goods and services would also encourage businesses to raise prices, and workers would expect their wages to rise in parallel to compensate them for the price rises. This would raise business costs, forcing them to lay off workers. Thus, unemployment would rise to its natural level – but inflation would stay high, potentially increasing dollar exchange rate volatility and currency risk for businesses.
Today, the Philips Curve in use is technically known as the “expectations-augmented Phillips curve,” and says that the more quickly workers adjust their wage expectations to rising inflation, the more quickly unemployment returns to its natural level. When unemployment is its natural level, inflation stops rising, the economy is running at full potential and the dollar exchange rate is stable.
The expectations-augmented Phillips curve underpins the Taylor rule, which is one of the principal equations the Fed uses in monetary policy decisions. The Taylor rule tests actual against potential output to determine what the right level of interest rates is for economic conditions, and hence it influences the exchange rate. The point on the expectations-augmented Phillips curve where inflation stops rising determines “potential output.”
Unemployment rose sharply in the Great Recession, peaking at 10 percent in October 2009.7 Since then it has steadily fallen, to 3.8 percent by February 2019, its lowest level since the late 1960s.8 But unlike the 1960s, inflation did not rise as unemployment fell, and the dollar exchange rate actually rose as the labor market strengthened. An explanation given by Lael Brainard, Federal Reserve Governor, is that people now expect inflation to be low, so don’t demand pay rises.9 Wage growth has indeed been subdued by historical standards since the crisis, running at below 3 percent until late 2018.10 Expectations of low inflation could partially explain the strength of the U.S. dollar exchange rate in recent years, since low inflation makes U.S. dollar-denominated assets attractive to investors.
However, there are alternative explanations. One is that the growth of temporary, casual, part-time and flexible contract work after the crisis led to many people being under-employed (doing fewer hours’ work per week than they want to), and others dropping out of the workforce because they did not think they could find employment – these are known as “discouraged workers.”11 Under-employment and labor market exit were not measured in unemployment figures but nevertheless exerted considerable downwards pressure on pay. However, under-employment is diminishing as the economy improves, and people who dropped out of the workforce are now returning. As a result, wages are now growing more strongly.12 This has yet to be reflected in higher inflation, but if and when it does, the U.S. dollar exchange rate could fall. So, the Phillips Curve relationship could still hold if the definition of “unemployment” is widened to include under-employment and discouraged workers.
Poor productivity may also explain the apparent failure of the Phillips curve relationship since the crisis. Since wage rises are typically associated with higher productivity, poor productivity may be helping to keep wage growth depressed, inflation low and the dollar exchange rate strong.13 Also, workers could simply have gotten used to low wage growth. This could all change if productivity improved.
Some economists say that the natural rate of unemployment could now be permanently lower than before the crisis. This could be due to changes to welfare benefit entitlements—for example, most states now have limits of six months or less for unemployment benefit claims, after which people must either find work or drop out of the workforce.14 Other reasons might include higher educational attainment and better matching of workers to jobs, both of which make it easier for people to find and retain jobs.15
Additionally, structural changes in the economy may have permanently weakened the relationship between pay and unemployment. Weaker labor unions and the dominance of very big employers mean that workers tend to have less bargaining power than in the past.16 Evidence from the U.K. suggests that in some industries, a government-imposed minimum wage has become the “going rate” for most jobs, creating little opportunity for workers to seek higher pay.17
Weakening of workers’ rights and the rise of the so-called “gig economy” create permanent job insecurity, making workers reluctant to seek higher pay even when unemployment is low. Technological advances also weaken the link between pay and unemployment, for example, because workers fear that if they ask for higher pay they will be replaced by robots. Additionally, today’s fluid FX markets and hedging instruments help businesses to manage the FX risk inherent in global supply chains, making it easier for them to relocate production and/or rely on foreign suppliers. This has helped to create global labor markets in which the “going rate” is effectively the lowest labor cost in the world plus the costs of maintaining supply chains.18
The economist Roger Farmer disputes the existence of a long-term “natural” rate of unemployment. He says that the unemployment rate is determined by psychological factors—what the famous economist John Maynard Keynes called “animal spirits”—and varies considerably. In his analysis, he replaces the Phillips Curve with a “belief function,” in which expectations of future GDP drive the decisions of firms and households. In simple terms, Farmer’s model says that if people believe future GDP will be low, it will be low, and unemployment will remain high.19 The implication is that there may be little relationship between unemployment, inflation, and the exchange rate.
The Phillips Curve relationship between unemployment and inflation appears to be weaker than before the crisis. This may be a temporary problem arising from the slow recovery since the Great Recession. However, long-term structural changes such as globalization, the advance of technology and the changing nature of work may render the Phillips Curve less useful than it has been in the past. A weaker Phillips Curve makes it harder for the Fed to judge what the full potential of the economy is and set interest rates that move towards full employment, target inflation, and a stable exchange rate. It therefore creates uncertainty for businesses looking to manage their financing costs and exchange rate risks.
With 17 years’ experience in the financial industry, Frances is a highly regarded writer and speaker on banking, finance and economics. She writes regularly for the Financial Times, Forbes and a range of financial industry publications. Her writing has featured in The Economist, the New York Times and the Wall Street Journal. She is a frequent commentator on TV, radio and online news media including the BBC and RT TV.
1. “NAIRU (long-term)”, FRED https://fred.stlouisfed.org/series/NROU
2. “Personal Consumption Expenditure Excluding Food and Energy,” FRED https://fred.stlouisfed.org/series/PCEPILFE#0
3. “Understanding the Disconnect between Employment and Inflation with a Low Neutral Rate,” Brainard https://www.federalreserve.gov/newsevents/speech/brainard20170905a.htm
4. “British Pound to U.S. Dollar Spot Exchange Rates for 1975 to 2019 from the Bank of England,” PoundSterlingLive https://www.poundsterlinglive.com/bank-of-england-spot/historical-spot-exchange-rates/gbp/GBP-to-USD
5. “Inflation rate in 1976,” UK Inflation Calculator http://www.in2013dollars.com/UK-inflation-rate-in-1976
6. “Unemployment statistics from 1881 to 1995,” UK statistical service (Terence Bunch) http://terencebunch.co.uk/articles/globalisation-the-united-states-empire-the-rise-of-thatcherism-and-the-uks-descent-into-dependency/unemployment-statistics-from-1881-to-1995-the-government-statistical-service-uk.pdf
7. “The recession of 2007-9: BLS Spotlight on Statistics,” Bureau of Labor Statistics https://www.bls.gov/spotlight/2012/recession/pdf/recession_bls_spotlight.pdf
8. US Civilian Unemployment, FRED https://fred.stlouisfed.org/series/UNRATE/
9. “Understanding the Disconnect between Employment and Inflation with a Low Neutral Rate,” ibid.
10. “Average Hourly Earnings of Employees: Total Private,” FRED https://fred.stlouisfed.org/series/CES0500000003#0
11. “Persons not in the labor force and multiple jobholders by sex, not seasonally adjusted,” BLS https://www.bls.gov/webapps/legacy/cpsatab16.htm
12. “Average Hourly Earnings of Employees: Total Private,” FRED, ibid.
13. “The labor market is booming – why aren’t your wages?,” Brookings, ibid.
14. 2019 Maximum Weekly Unemployment Benefits By State,” Saving to Invest https://www.savingtoinvest.com/maximum-weekly-unemployment-benefits-by-state/
15. “The Phillips Curve: lower, flatter or in hiding?,” Cunliffe, https://www.bankofengland.co.uk/-/media/boe/files/speech/2017/the-phillips-curve-lower-flatter-or-in-hiding-speech-by-jon-cunliffe.pdf
16. “The labor market is booming – why aren’t your wages?,” Brookings https://www.brookings.edu/blog/up-front/2018/10/26/the-labor-market-is-booming-why-arent-your-wages/
17. “Pay In The Adult Social Care Sector,” Skills for Care https://www.skillsforcare.org.uk/NMDS-SC-intelligence/Workforce-intelligence/documents/Pay-in-the-adult-social-care-sector.pdf
18. “The Phillips Curve: lower, flatter or in hiding?,” Cunliffe, ibid.
19. “Keynesian economics without the Phillips curve,” Farmer http://www.rogerfarmer.com/rogerfarmerblog/2017/10/1/keynesian-economics-without-the-phillips-curve