In its recently released 2016 Guide To Retirement,JP Morgan takes a fine-tooth comb tojust about every imaginable variable impacting retirementreadiness.

But it is the report's “retirement savings checkpoint” where JPMorgan introduces the wild card that may turn conventionalretirement preparation thinking on itshead for the decade going forward.

In modeling ideal savings levels relative to age andincome (a 40-year old making $50,000 a year should have 1.2 timesthat in savings; a 60-year old making $100,000 should have 7.3times that in savings), JP Morgan presumes 6.5 percent annualizedreturns on portfolios for pre-retirees, and 5 percent returns forretirees.

Those assumptions are markedly lower than historical returns.Between 1970 and 2014, the annual compound return on large-capstocks was 10.5 percent, and for bonds it was 7.9 percent,according to data cited by Charles Schwab.

Like others, JP Morgan has tempered return projections, thoughnot nearly as much as some retirement providers and investmentluminaries.

Last October, Jack Bogle, Vanguard's founder, told Morningstar'sChristine Benz that annualized returns on stocks could be as low as4 percent for stocks in the next decade, and 3 percent for fixedincome.

That would create a 3.5 percent nominal return on a balancedportfolio. Factor in inflation, and then all fees, and pre-retireesstand the chance to actually lose money on their portfolios overthe next decade.

“You're talking about a really tough decade for equityinvestors,” Bogle told Morningstar. “Predicting at 7.5 percent or8.5 percent return is just silly.”

Bogle acknowledged he could be wrong. But he's certainly notalone in his bearish projections.

David Blanchett, head of retirement research at MorningstarInvestment Management, told BenefitsPro that he routinely talks toadvisors who prefer to use historical return norms, which wouldapply the historical return on U.S. government debt of 5 percent,well below the 2 percent range Treasuries now yield.

That approach does not square with reality, as Blanchett seesit.

“It's bad news for a lot of people. Many pre-retirees are goingto have to save more and retire later—that's just the way it is,”he said.

“I think returns going forward are going to be lower thanhistorical averages,” added Blanchett. “The U.S. has pretty muchbeen the best performing market for stocks in the world over thelast 115 years, and I don't think it's realistic to assume thisrelative outperformance is going to continue.”

Blanchett cautions that return estimates have to be adjusted fordifferent time horizons. Assuming the next decade's assumptionswill hold true beyond that period is questionable, he says.

That matters for today's pre-retiree that is hoping to retire in10 years; a 55-year old still has an investment horizon aspotentially long as another 35 years, or more.

Blanchett said Morningstar is projecting a 6.5 percentannualized nominal return over the next 10 years—inflation, taxesand fees will take their share.

Morningstar expects U.S. large cap stocks to return 5.4 percentnominally in the next decade, while U.S. small caps will do better,at 6.8 percent. The firm's fixed-income expectations are in linewith Bogle's—around 3 percent.

International stocks will post 8.1 percent nominal returns overthe next decade, according to Morningstar.

While that would certainly be welcomed, Blanchett cautionspanicky pre-retirees that may be forced to play catch up during adecade of tepid returns.

“Lower returns and low interest rates will definitely affectallocation models, but you are not going to be able to re-allocateyour way out of a savings shortfall,” he said.

Other service providers are basing retirement readiness modelson projections in line with Bogle's conservative estimations.

Fidelity assumes 3 percent annualizedreturns in its modeling, a conservative return used toaccount for market volatility and unknowns, like health careinflation in coming decades.

To Blanchett, the reality is clear, irrespective of variances inreturn projections.

“Today's advisors are going to have to use lower returnprojections in their modeling.”