The Conference Board, a nonprofit for economic research, tracks 11 predictive measures of future economic activity in its Index of Leading Economic Indicators. The LEI purports to forecast the economy over the coming three to six months. The individual components include data on unemployment, the direction of the stock market, consumer and business sentiment, and manufacturing activity. Unfortunately, the LEI is somewhat unreliable as a forecaster and often misleading.

We examined the record of the LEI (and its components) over the eight recessions and nine sudden market declines of 15% or more since 1960. The good news is that the LEI and many of its components have had a near-perfect record in anticipating recessions. The most reliable indicators have been the shape of the yield curve, business and consumer confidence, durable-good purchases and housing starts, and the health of the labor market. These measures have also correctly signaled stock-market downturns. (The biggest exception is consumer spending, which has risen before nearly every past recessions, falling only after the recession starts.)

Yet despite the LEI’s fair record, there are good reasons to doubt that any economic statistic can reliably predict when a downturn will occur. Since 1958, even when the indicator was correct, the lead time between its turning negative and an economic slide has been as long as 18 months. Worse, there have been many false positives. Leading indicators have incorrectly forecast a downturn many more times than they correctly predicted recession or stock-market decline. In fact, the most accurate indicators have the highest incidence of false positive signals. A signal that often predicts recessions that don’t happen is more misleading than helpful. As the economist Paul Samuelson once quipped, “The stock market has predicted nine of the last five recessions.”

Today, the auguries are generally favorable. Stocks and the labor market have performed well, and the yield curve is sloping upward again. But business confidence has declined over the past three months through November, based on uncertainty about trade and global politics.

Our view is that the best course of action is to be agnostic about future economic activity and the direction of the stock market. We subscribe to the wisdom of John Kenneth Galbraith, who once said, “There are two kinds of forecasters: those who don’t know, and those who don’t know they don’t know.”

If accurate forecasts of the economy and the stock market are impossible, how should investors assess their portfolios at year-end? There are three steps every investor should take, none of which require futile attempts to forecast the future.

First, ensure that their asset holdings are broadly diversified. Hold internationally diversified equities and real assets such as real estate that will benefit if growth continues or inflation accelerates. Also hold safe assets, such as fixed-income securities, that would balance the losses in a recession.

Second, maintain the balance of their portfolios to suit their retirement timeline and risk tolerance. If equity holdings increase so that the risk level of the portfolio is too high for comfort, for instance, it may make sense to sell off equities and reinvest in safer asset classes. Rebalancing always reduces risk and in volatile markets can increase returns.

Third, be sure to harvest tax losses and keep costs minimal. Losses should be realized on assets that have declined in price. It’s possible to deduct the loss on any asset sold even if one is reinvesting the proceeds in a different asset class for balance. Net losses, up to a certain amount, can be deducted from income taxes. Low- or zero-cost index funds should be your favorite investment vehicles, and portfolio management costs should be minimized. The greater the costs and fees you pay, the lower your returns. As John Bogle used to say, “You get what you don’t pay for.”

This op-ed originally appeared in the Wall Street Journal on December 30, 2019