Watch out, traders, the real fed funds rate is about to turn positive for the first time since 2008. Or is it? That depends on which inflation measure you use. The bad news is that deflating the Fed funds rate by the widely watched core PCE deflator suggests that policy is no longer particularly accommodative. The good news is that you’ll only have to live in this brave new world for a couple of weeks, because without food and energy you’ll quickly expire.

Comparing the policy rate with headline CPI, on the other hand, suggests that the Fed could hike rates two more times and retain a negative real funds rate. While this might seem like an argument over semantics, the historical record suggests that investors can improve their forecasting performance by choosing the right inflation measure to follow.

Economists like to focus on measures of core inflation because it provides a more accurate picture of future inflation trends than headline readings. In the real world, however, food and energy are basic necessities — and price swings influence household finances and spending. So which measure should markets focus on?

While there is of course a broad correlation between headline and core inflation readings, their trends diverge surprisingly often. It turns out that following core inflation is the right call for markets — but not the measure the economists prefer.

If the real policy rate matters for equities, we would expect to see a negative correlation between it and subsequent stock market returns. In fact, this is exactly what happens. Since 2000, the correlation between the real Fed funds rate and subsequent 1-year return of the S&P 500 has been negative, regardless of which inflation measure you use to calculate the real funds rate.

However, it’s clear that using core measures yields a stronger relationship; over the full sample, both core PCE and core CPI have a correlation of roughly -0.5 with subsequent equity returns. Over the last couple of years, however, using the funds rate deflated by core CPI yields a correlation of -0.82, while the core PCE equivalent is just -0.13.

What about fixed income? Inflation rates and the yield curve are both stationary series, and thus an apt focus of comparison. Interestingly, over the past couple of years the curve has ignored trends in inflation. That makes sense: inflation rose in 2016 and fell this year, but the curve has just kept on flattening.

Over time, however, the relationship with core CPI is clearly stronger than with other measures. Since 1995, the slope of the 2s10s Treasury curve has a correlation of -0.66 with core CPI but just -0.40 with core PCE. Correlations with headline inflation are even lower.

Now, to some extent this might reflect a timing issue. Core CPI is typically released a couple of weeks before the PCE measure and thus informs the inflation narrative. Still, that’s useful information to know. Either way, if and when the Fed funds rate exceeds the core PCE deflator, the numbers suggest that you shouldn’t dress for the equity bull market’s funeral quite yet.