No one really knows how to predict recessions. Despite the massive financial returns that would accrue to anyone who had a good way of spotting downturns before they happen, macroeconomic forecasters are unable to do this with any reliability. And if professional forecasters can’t succeed in the high-stakes world of the financial industry, the chances of a pundit such as myself to see a recession coming seem fairly remote.


Still, it’s fun to speculate. With the extreme caveat that no one really knows what they’re doing when it comes to predicting the economy, let’s take a look at some potential signs of a recession in the offing.


At the moment, the economy is doing well. There are more job openings than job seekers — a first since 1970. The prime-age employment-to-population ratio, probably the best indicator of the health of the job market, is back to 2005 levels. Exports and business investment are both robust.


To some, this healthy performance might seem like a sign that a turning point must be drawing near. Writing in Forbes, Raul Elizalde says that a recession “seems to follow” whenever weekly unemployment claims go below 300,000. But if that were true, the U.S. would have had a recession back in 2015:


Meanwhile, statistical analysis of post-WWII U.S. recessions, though admittedly suffering from a small data sample, doesn’t show any indication that economic expansions die of old age.


Elizalde also suggests the Treasury term spread — the difference between long-dated Treasury yields and short-dated yields — as a leading indicator of recessions. Research by staff economists at the Federal Reserve Bank of New York supports the notion that the term spread is the best method of forecasting downturns 12 to 18 months in advance. When long-term rates fall relative to short-term rates, it can signal that investors see higher risk in the short term. The extreme form is when the yield curve becomes inverted, or when short rates are higher than long rates. Term spreads now are fairly low, but are not yet in negative territory.


Meanwhile, the Federal Reserve is slowly raising interest rates. Median forecasts by central bankers project the federal funds rate, now at 1.7 percent, to rise to 2.9 percent by the end of 2019 and 3.4 percent by the end of 2020. Typically, rate hikes precede recessions:


In a historical sense, 2.9 percent or even 3.4 percent is not high, but it’s noticeable that every pre-recession interest-rate peak since the early 1980s has been lower than the last.


Another metric to look at is oil prices. Economist James Hamilton has hypothesized that sharply rising oil prices lead to recessions, noting that every such spike from the 1970s through 2008 was followed by an economic slowdown. In 2011, shortly after Hamilton made his argument, the pattern was broken, as a large oil price increase failed to send the economy back into the dumps. Nevertheless, oil prices may be an important harbinger of recession. Prices are rising, having gone up by about a third since last year:


That’s a worrying sign, but if the economy could weather $100 a barrel oil a few years ago, it seems like there’s a good chance it could handle $70 oil now.


Another potential cause of recession would be policy mistakes. President Donald Trump’s trade war has business economists feeling jittery — recently, two-thirds of the members of a National Association for Business Economists panel predicted a recession by 2020, and blamed the trade war as the primary culprit.


One final possible cause of recession is the financial sector. Fed economists have found that when credit-market sentiment is high — indicated by low spreads of risky bonds over safe bonds — it tends to predict a reversal and a credit crunch two years later. Credit spreads now are at fairly low levels, having come down quite a lot from 2016:


If the pattern holds, a tightening of lending might be in the cards for 2020.


Taken together, the tea leaves generally point in the same direction — there’s potential trouble ahead. Falling term spreads, low credit spreads, rising oil prices and rising interest rates could point to a recession in late 2019 or 2020 — just in time for the next presidential election, which depending on your perspective is either great or worrying.


Of course, given the difficulty of predicting recessions, that forecast must be regarded as a highly uncertain one. Just remember, don’t bet on macroeconomic indicators.


Noah Smith is a Bloomberg Opinion columnist. He was an assistant professor of finance at Stony Brook University, and he blogs at Noahpinion.