As congressional Republicans prepare to pass their tax bill, the Federal Reserve is about to say goodbye to Janet Yellen as chair. She's had a good run: The United States and the world recovered from the financial crisis; steady, if unspectacular, growth resumed. Yet, now the Fed is in an unusual spot as Jerome Powell takes over. Having proved itself a creative and indispensable crisis manager, it now risks repeating the worst mistakes of central banks past. Yellen's last major act, to raise short-term interest rates again this past week, shows that the Fed still cleaves to rigid models even as reality refuses to conform.

Yellen's explanation, as with other recent hikes, holds that the Fed expects higher gross domestic product growth, a strong labor market and, hence, higher inflation in the coming years. Historically speaking, these trends have always been linked. But in recent years, something has changed. GDP growth now exceeds 2.5 percent annually. The official unemployment rate is at a generational low of 4.1 percent. But inflation? Despite forecasts from every stripe of economist, it simply hasn't materialized. This past week, the Bureau of Labor Statistics reported that inflation is running at 1.7 percent, lower even than the slow pace that had been expected by the Fed and most economists at the beginning of the year.

And so the Fed is dispensing medicine before the patient is ill. It plans to raise interest rates three more times in 2018 — above the current rate of inflation. It hopes to cool down the economy and prevent purchasing power from falling. But it has been anticipating since the financial crisis that this would happen, even though inflation has rarely exceeded 2 percent. The models, which worked in the past, dictate that the Fed should act prophylactically. Perhaps. But it is more than possible that the new economy demands that the Fed confront structural changes rather than assuming that at some point soon, enough inflation will return and everything will revert to the desired norm.

The Fed acted nimbly in the last crisis: It slashed short-term rates aggressively and then began a bold and untested policy of quantitative easing. Some say those moves created shadow inflation for the wealthy, in things such as stock prices and even art. But measured inflation didn't go up, and economic activity did. Audacious as the Fed was then, however, its institutional DNA (like that of all central banks) is to stick to models that have been developed meticulously over decades. One of the main critiques of central banks past is that they were staffed by economists and policymakers more attached to the perfection of their models than the messiness of reality, willing to let the world burn rather than sacrifice perfect theories, no matter how untenable. The "lords of finance" so ably described by Liaquat Ahamed in his work on the causes of the Great Depression were the progenitors. But their heirs in the 1970s, whose models suggested a measured approach unsuited to spiraling stagflation, and in the mid-2000s, when the soaring housing bubble was seen as irrelevant to inflation, proved equally wedded to theory at the expense of pragmatic policy.

Almost all the relevant economic models today suggest that inflation is about to bust out, even if it hasn't. Yellen conceded this point in her news conference Wednesday, noting: "We still expect inflation will move up and stabilize around 2 percent over the next couple of years. Nonetheless, as I've noted previously, our understanding of the forces driving inflation is imperfect." The Fed certainly deserves some benefit of the doubt, given how well it navigated in 2008-2009. But there's also a decent chance that the Fed is wrong in its inflation expectations, is wrong about interest rate policy and is steering the monetary ship in the wrong direction by raising interest rates anyway.

The expectation of inflation is based on cycle after cycle in the 20th century. Time and again, as economic activity as measured by GDP (or GNP, gross national product, for many of those years) accelerated, so did wages. As wages picked up, so did inflation and inflation expectations. Workers on the whole, especially when unions were strong, were able to demand a greater share, and as their salaries increased, so did spending power, and soon enough the costs of goods and services. Ideally, the Fed was supposed to note the pickup in activity and raise interest rates — the idea is that by increasing the cost of capital, higher rates act as break on spending, leading, in turn, to slower growth — before inflation became too potent.

Why has that cycle stalled in the past decade? One theory is that the financial crisis was so deep that it has taken years of easy (or "cheap," if you prefer) money to climb back to anything approaching normal. Another holds that, with all the novel measures the Fed took to halt the Great Recession, including quantitative easing, much higher interest rates will be needed to return to the status quo. A third posits that although the official unemployment rate is low, there are still millions who are underemployed or out of the workforce, and hence employment in America isn't quite as rosy as it appears.

Those explanations suggest that, while inflation has not yet returned, it's only a matter of time, so it's prudent for the Fed to take action now, gradually, while it can. But other theories posit that something has changed permanently and structurally. The McKinsey Global Institute has made waves with its bold predictions of an increasingly robotic future that will make many jobs obsolete. And it doesn't take an elite consultancy to notice that technology is creating massive deflation of goods and services: A smartphone that retails for less than $1,000 does the work of equipment that cost 10 times that a decade ago, and food and clothing (once a considerable portion of a typical family's budget) continue to decline in price. What's more, the disparity between skilled workers, whose wages are rising, and workers with fewer or outdated skills means there can be inflation in urban hubs like New York and San Francisco and deflation everywhere else, which would make broad-brush Fed policy into a bazooka, helping and harming simultaneously. The Fed staff surely knows that, but the bank cannot set regional monetary policy, only national.

If we're simply in a slow-motion version of a familiar cycle, then the Fed's current approach will prove just fine. But if there's a deep structural change in our economic system — brought on not just by technology but by the global emergence of a multibillion-strong middle class and by supply chains and business stretched across the globe, no matter how much the current occupant of the White House may disapprove — then using old models to fight nonexistent inflation will wind up being a potentially costly miscalculation, prematurely putting the brakes on an economy that's slowly healing.

The task for the Fed and the world's other central bankers is hardly enviable or easy. No institutions or their stewards abandon old models and plunge into the new without being forced to by circumstances. The familiar is almost always easier to defend than the novel, even if that proves the wrong path in hindsight. There's a reason people always fight the last war.

For now, however, there is no crisis, which gives bankers and economists the time and space to consider what to do if the old model is no longer applicable. In the recession of 2008-2009, urgency dictated creativity. Today, alas, calm is fostering complacency.

Karabell is the head of global strategies at Envestnet, a financial services firm, and the author of "The Leading Indicators: A Short History of the Numbers That Rule Our World."