Do you think access to faster and more information has changed the way we form expectations? Or does technology have no effect on the way people adapt to policy? If so, what effect does technological change have on the Phillips curve?
Beginning in the 1980s with 24-hours news on television and then spreading across dozens of such channels, spilling onto the then nascent internet in the early part of the 1990s and finally beginning to fracture into increasingly personalized bubbles of information with the reemergence of patently partisan news sources (a move back towards the standard news landscape from the founding of the republic until well into the 20th century), information and its delivery has become a major commodity driving our economy. The big bang over the last decade which has spewed forth all manner of spin being put on every facet of news and information by myriad think tanks, a situation mimicked by millions of inwardly-focused, fetishistic blogs and websites. This has perhaps had beneficial effects for those dedicated enough to dig through dozens of sources in forming opinions on any given event or issue. But for many more, the world has turned into a segmented labyrinth of desired viewpoints in which one writer’s square is another’s circle.
It is from this cauldron that we can have a world in which one person thinks we are on the verge of deflation while the next thinks that hyperinflation is going to break out first thing next week. Such wildly opposing views make the idea of any kind of homogeneous formation of expectations hard to map out. The graph below shows both the University of Michigan Consumer Sentiment survey and the BLS Consumer Price Index (less food and energy).
If there is a relationship to be gleaned here, it seems to be one of increasing divergence between actual inflation and expectations. If anything, consumer expectations nearly always overstate actual inflation. It may be argued that the less volatile metric of the CPI, less food and energy, is a poor comparison with the U of M survey, since respondents would typically feel the full volatility of swings in food and energy as part of forming the expectations reflected therein. Nonetheless, if we are attempting to analyze whether increasing information availability helps expectations adapt to policy, then the relevant point is whether such information is helping to form expectations consistent with the metrics used in policy formation (namely measures such as the CPI minus food and energy). Much evidence points to decreasing consensus between legislators, central bankers, investors, pundits and the general public in this regard.
No clear evidence is apparent on a relationship between the increased information landscape and the durability of Phillips curve principles. Many (most?) economists have largely discarded the Phillips curve as anything more than a sort of teaching tool to understand issues related to the principal of the neutrality of money. Many heterogeneous factors have played into the relationship between unemployment and inflation over the last three decades. To name a few, the profound decline in the power of organized labor to set wage norms, massive increases in income disparity allowing an affluent minority to consumer an outsized portion of output through good times and bad, and the blowing up of a series of asset bubbles allowing for consumption increases through growing leverage and/or indebtedness on the part of many Americans even in the face of declining real wages and employment security. These types of complex relationships defy a monolithic causal connection between inflation and employment.
Some economists are even beginning to argue that the era of secular low inflation and accompanying low interest rates has led to a reduction of the effectiveness of the tools of monetary policy, as the zero lower bound is always looming just 300 or 400 basis points below the typical rates used to regulate economic activity over the last few decades and once it is reached, monetary policy is largely stuck with highly irregular tools such as quantitative easing and other potentially mixed effect policies. This analysis leads to arguments that the effectiveness of monetary policy would be well served by a higher level of inflation and interest rates, but the current political environment is unlikely to yield serious considerations of such arguments.
In conclusion, the vast explosion of accessible information over the last few decades has had little impact on the relevance of the Phillips curve, which was of limited relevance in the first place (both in its original form and within the modified Freidman/Phelps paradigm). Both versions assume the primacy of the relationship between wage growth and price inflation, a relationship that has not shown a strong correlation since the 1980s. Newer models of inflation analysis, such as Gordon’s Triangle, which explicitly recognize supply shocks and built-in inflation as well as the traditional Phillips curve paradigm, seem to point a better way of understanding the complex events that form and sustain inflationary expectations, and allow for a much more dynamic view of the web of relationships between inflationary expectations and actual inflation.