The comptroller of New York State, Thomas P. DiNapoli, recently announced that the New York State pension fund will divest itself of many fossil fuel stocks within five years. He said, “investing for the low-carbon future is essential to protect the fund’s long-term value.” He had resisted such a strategy for many years because of his fiduciary duty to fund the retirement benefits of state and local public sector workers.

Investment restrictions are increasingly common. 26 percent of all U.S. professionally managed assets—worth $12 trillion—are governed by investing limitations; $3 trillion have fossil fuel restrictions.

And some claim that the tradeoffs Di Napoli worried about have disappeared: you can sell fossil fuel assets and improve returns. The Rockefeller Brothers Fund claims to have reduced its exposure to fossil fuel companies and increased its returns relative to a benchmark investment portfolio.

How should we evaluate such claims? What can disinvestment achieve?

If investment markets are efficient, then all known factors affecting the future returns of assets are already incorporated into the values of stocks and bonds. Thus, if investors believe that future climate change policy will reduce the demand for and the future returns from fossil fuel investments, that knowledge already has been incorporated into the relevant stock and bond prices. New York State’s (or any other investor’s) decision to reduce fossil fuel investments over the next few years cannot prevent investment losses of this type because they have already occurred and will recur in the future if additional negative information arises.

What if fossil fuel demand continues for decades and fossil fuel investments generate profits for decades? How does a policy of shunning fossil fuel assets affect the returns of those assets and the other assets that remain in the portfolios of those who divest?

Before divestment occurs, assume that all investors hold the portfolios of stocks and bonds that they prefer. Many are index investors who own the entire market. Others weight assets differently because they have beliefs about future earnings of sectors or companies that are not incorporated in current asset market prices.

Now assume that some holders of fossil fuel assets want to sell them to signal their moral disapproval and/​or force the companies to change behavior. The first question is who will buy the shunned assets and at what price? The only buyers are individual investors or managed mutual funds. And both already were satisfied with their portfolio composition. To induce either to (in effect) “overweight” fossil fuels in their portfolios, the price of fossil fuel assets must fall, which compensates for the future “thinner” resale market for such stocks because of “shunning.”

The result is that whatever price would be consistent with the cash flows generated by fossil fuel stocks if they were “normal” firms must now be discounted to induce anyone to own them because the stigma associated with the industry implies that the future ability to resell the stock is lower. One receives the same expected future cash flows for a lower price to compensate investors for the “thinner” resale market.

What happens to the prices of those investments bought by those who shun (and thus sold) fossil fuel stocks? Using the same logic as before, the individual investors who purchase the fossil fuel stocks at a discount simultaneously sell other stocks at a premium to those who shun fossil fuel stocks.

By lowering the price on existing investments, disinvestment does reduce investment in future fossil fuel projects. Investors’ willingness to pay for estimated future cash flows from new fossil fuel projects decreases to be consistent with the lower price for current cash flows from existing investments. Thus, the set of projects for which investment covers costs and produce sufficient cash flows to induce investment is now smaller. Stated differently, lower prices for fossil fuel assets mean a higher cost of raising funds for new investments because, at a lower price, fossil fuel firms have to issue more shares to raise a given amount of funds, which is costly to existing owners because it dilutes their existing ownership.

Is there evidence consistent with this discussion of theory? A recent paper examines the returns of all public U.S. companies from 2005 through 2017. After controlling for variables that predict investment returns, firms with higher CO₂ emissions generate higher returns. A one-standard-deviation increase in the level and change of total emissions leads to 1.8% and 3.1%, respectively, increases in annualized returns.

To summarize, disinvestment decreases the price of disfavored stocks and increases the price of all other stocks. As long as all other factors remain constant, fossil fuel stock returns increase and all other stock returns decrease. But the silver lining for those who advocate disinvestment is a higher cost of capital for and thus fewer future fossil fuel projects. Thus, disinvestment does promote the objectives of its advocates; but claims that doing so is costless do not stand up to theory or evidence.