Planning our financial futures has never been easy. But today, more than ever, we need to take the task seriously. We also need better tools to do it.

We believe the "dismal science" offers a better solution than the standard prescriptions of conventional financial planning. The solution from economics is called "consumption smoothing." The idea has been around for nearly a century, tucked away in academia. It had no practical use because the computing power necessary to do it simply didn't exist. Today, that computing power is available in a laptop. So is the first generation of software.

The software, ESPlanner, was developed by one of the authors, Larry Kotlikoff. Scott Burns has no financial stake in the software. As a journalist, Burns has been interested in consumption smoothing since his first book on personal finance was published in 1972. Our purpose, here, is to explain the idea of consumption smoothing and show how and why it is superior to conventional financial planning.

Most conventional financial planning is based on simple time value of money exercises, projecting and discounting flows of savings and investment. Much of it is stuff that you could do using a handheld financial calculator, first introduced by Hewlett Packard in 1970. Needless to say, there are more sophisticated tools today, but even the best of the snazzy Monte Carlo simulation tools start from the same flawed premise--- they pick a retirement income target by rule of thumb. Then they do time value of money calculations to tell you whether you'll have the necessary income at the appointed time.

A 40 year old couple with combined earnings of $80,000 a year and receiving inflation-only raises, for instance, might be told that they would be earning $148,824 a year by age 61 and that they would need to replace 85 percent of that at retirement, or $126,500 a year. To produce that income they would need a nest-egg of 20 to 25 times that income, or at least $1,339,400. Another calculation would tell them how much they needed to save each year to get there. Assuming an 8 percent return, they would need to save over 20 percent of income if they had no savings to start. They would need about 8 percent of income if they had starting assets of $150,000.

The intimidating savings rate could be lowered if future Social Security benefits or a pension were considered, but other projects--- such as college educations for children--- are usually treated separately and call for additional savings. Basically, conventional software puts savings first. It leaves personal consumption as an uneven residual--- you get to spend what's left after all the savings projects have been fulfilled. As a consequence, personal consumption fluctuates and conventional software calls for annual savings most people simply can't do.

Consumption smoothing starts from a different place. It also requires massively more synthesis and computation. It uses dynamic programming to calculate a level personal consumption path each year until death. It smoothes your consumption through retirement by adjusting the amount you save or dis-save. It does this year by year, based on the changing realities of your life.

Consumption smoothing means not guessing a constant level of saving and then finding out, in retirement, that it's not enough. It also allows you to avoid a meaner fate--- enjoying a well financed old age after a deprived and pinched youth. The idea of consumption smoothing underlies everything economists have to say about saving, insuring and investing.

Until now, economists have had no practical way to convey consumption smoothing to the public. It's a "slow idea"--- difficult to get your head around. Determining how much households need to spend, save, insure and diversify to smooth their living standard is also an incredibly complicated computational problem. The short list of interconnected factors includes household factors like the number of children (if any) at home, earnings, taxes, housing plans, the economies of shared living, types of retirement accounts, mortgages, special expenditures, pensions, Social Security benefits and estate plans. And the problem has to be solved for all years, simultaneously.

Here's what consumption smoothing can do for our hypothetical couple, the Jays. Both are 40, married, live in Wisconsin, and make $40,000 each. They have a 6 year-old and a 10 year-old. They also have $150,000 in regular assets, a $300,000 house, and a $150,000, 20-year mortgage with a $1,000 per month payment and other commensurate housing expenses.

They plan to pay college expenses of $20,000 a year for each child (in today's dollars) for 4 years. They also plan to retire and start collecting Social Security at age 62. Being conservative, the couple invests only in TIPS — Treasury Inflation Protected Securities yielding 2 percent above inflation. They plan to do this in IRAs, saving $3,000 a year each or 7.5 percent of their gross earned income. They also plan to pay off their mortgage over the appointed period.

What does consumption smoothing tell them?

Although total spending changes over time, consumption smoothing delivers a stable real consumption of $20,197 a year per adult. It does this year in and year out. (This is the amount of spending each adult would need as a single person with no children to enjoy the same living standard he/she enjoys in the household. Shelter expenses, including mortgage payments, are a separate item because they may change their shelter.)

The actual dollar spending varies with the composition of their household. While both kids are at home consumption smoothing suggests $46,309 in annual consumption. It drops to $39,308 when the first leaves for college. It bottoms at $32,315 when both kids are out of the nest. All figures are in real dollars, adjusted for inflation.

Is that smooth consumption? Yes. Each figure takes into account the size of the household and the economies of shared living. While two can't live for the price of one, the accepted figure is that two can live for 1.6 times the cost of one. So a constant $20,197 per adult in a single person household becomes $32,315 for a couple. For this couple, consumption rises an additional $7,000 for each child at home. The graphic below shows the lifetime flow of expenditures.

As you can see, their consumption is a smooth $32,315 once the children are educated and no longer at home. Before that, their consumption is higher because their household is larger. What's most important, however, is that their adult living standard is constant at all ages even though their retirement income is about 50 percent of their pre-retirement earnings.

Lifetime Savings and Dissavings graph While the Jay's are adding a real $3,000 a year, each, to their IRA accounts, consumption smoothing maintains their consumption by changing the amount they save, or dis-save, each year. It adds to, or withdraws from, the regular assets they started with, beginning with $2,319 in dis-savings the first year. They massively dis-save ($9,500 to $17,155) a few years later while the kids are in college.

Once the kids are off the payroll, they have five years of heavy duty saving. During that time they put away $10,800 to $18,900 a year, just prior to retirement. Then, beginning at age 62, they dis-save every year. This is how consumption smoothing works out a smooth lifetime path for their adult standard of living.


Sounds like real life, doesn't it? We do what we can, when we can.

Between age 40 and 61 the Jays contribute $132,000 to their IRA accounts but dis-save a net of $51,700 from their other assets. Their net saving for retirement is $80,300. That's only 4.6 percent of their total earnings over the period. So consumption smoothing tells us that you can achieve a lifetime of smooth consumption, educate two children, and pay off a home mortgage for a fraction of what conventional planning usually suggests. Their retirement spending is $38,315 ($32,315 for consumption and $6,000 for shelter expenses). They also pay $144 for income taxes in their first year of retirement but pay much more in later years. Their retirement income replacement rate is about 50 percent of their $80,000 pre-retirement income.

That's massively lower than the conventional financial planning estimates of 70, 75, and 85 percent of final salary.

Yes, we've made the benign assumption that Social Security will be there for them. If we assume Social Security benefits will be lower their savings would need to be higher.

Can they do better? Of course! They can raise their lifetime consumption by taking some risk and getting a higher return on their savings. They could also take Social Security late rather than early. We'll cover some of their options in another column.

Sadly, conventional financial planning asks us to do this rocket science in our heads. Or it uses rules of thumb that don't work or simplify and omit so many interactions that recommendations are as useless as our guesses.

And that's a problem. Even mistakes of 10 percent, because they apply to all the potential years of retirement and survivorship, can lead to huge mistakes in saving and insurance recommendations. They can also lead to major disruptions — on the order of 30 percent — in our living standards when we retire or become widowed.

Sales vs. your needs

Financial guessing based on the retirement income replacement rate concept is ubiquitous. For large investment companies it works nicely as a marketing tool. Their web calculators suggest very high saving targets and offer simple and quick answers. The goal is to move us quickly from planning to purchasing. The investment companies aren't alone, however. Media sources use the same flawed and misleading method.

Hypothetical couple

Skeptical? Let's see what two popular online calculators would tell the Jays. Remember, ESPlanner has told them that a gross savings rate of only 7.5 percent, and a net savings rate of only 4.6 percent will help them educate their children and maintain a constant adult consumption standard throughout their lives. It also determines that to do this their retirement income replacement rate will be under 50 percent.

TIAA-CREF's Retirement Goal Evaluator recommends a retirement spending target equal to 80 percent of annual earnings. For our stylized couple, this target is 60 percent too high.

CNN Money's retirement calculator recommends a retirement spending target equal to 70 percent of annual earnings. They then offer a variety of portfolios and calculate the odds of meeting the stated goal, which is 40 percent higher than consumption smoothing indicates.

How can you check the results of a consumption smoothing black box? You simply verify that recommended annual spending entails a stable living standard per household member and that the household dies broke (with exactly zero assets). Given this, any higher spending path would drive the household into debt, and any lower path would have it leave money on the table.

Consumption smoothing doesn't blindly guess a future consumption target and then calculate the saving needed to pay for that guess. Instead it uses advanced mathematics and computer technology to calculate the household's highest sustainable living standard and the annual spending and saving needed to preserve that living standard. In short, it actually plans for households rather than make them plan for themselves.