Donald Trump’s campaign headquarters in August in New York City. (Alison Elkin/Bloomberg)

For months now, the Monkey Cage has been running an election forecasting tournament with Good Judgment. Recently we asked forecasters this question: “As of Nov. 1, 2016, how much money will the Center for Responsive Politics report that Donald Trump’s campaign committee and outside groups have spent on his campaign?”

About three months ago, the forecast for Trump was remarkable: There was a decent chance that Trump would spend less, in inflation-adjusted dollars, than did John McCain in 2008. McCain, let’s recall, accepted public funding for the general election and therefore had his fundraising and spending limited (unlike his opponent, Barack Obama, who did not accept public funding).

The forecast now is even more pessimistic, even as Trump reports $90 million in fundraising in August. Forecasters now give Trump a 73 percent chance of spending $250 million to $500 million.

Given that Trump’s total as of Aug. 22 was about $161 million, including fundraising by outside groups, his campaign could ultimately beat McCain’s haul ($368 million). But Trump is still far behind Clinton and her affiliated outside groups, which have raised $435 million.

Indeed, Trump’s haul is likely to be a historical anomaly, especially in an era when presidential candidates do not take public financing in the general election and when independent groups raise and spend hundreds of millions of dollars on behalf of the presidential candidates they support.

Now the interesting question becomes: How much will this matter? In this era of partisan polarization, does any major party nominee have a floor of support, no matter what they spend? Can Trump rely on free news coverage, or even the Clinton campaign’s own perceived missteps, to close the 5-point polling gap between them?

Perhaps the best way to think of Trump’s anemic fundraising is with a term he might understand as a casino owner: a gamble.

The best evidence suggests that the work of presidential campaigns does produce votes. When one candidate can out-advertise the other, the polls appear to move in that candidate’s direction. See, for example, Lynn Vavreck’s and my findings in our book on the 2012 election. An ad advantage may also translate into real votes. See this article by political scientists Michael Franz and Travis Ridout on Obama’s ad advantage in 2008. Moreover, the ground game does increase turnout among a candidate’s fellow partisans.

Altogether we might not expect the ads and ground game to add up to more than a few points of vote share. But can Trump afford to give up those few points, especially when he’s currently the underdog? That seems like a gamble Trump is increasingly willing to make.