Here's how: by simply directing the Census Bureau to produce an honest measure of poverty, as opposed to the obviously fake and deceptive measures that they currently produce.

In September the Census Bureau came out with its annual report on Income and Poverty in the United States, covering 2014.  The so-called "official poverty rate" was reported as 14.8%, allegedly representing some 46.7 million people living "below the poverty line."  The 14.8% rate actually was an increase from 2013's 14.5%, although the Report admits that the increase was not "statistically significant."

Still, with a trillion dollars or so of spending on anti-poverty programs at federal, state and local levels each year, and with an economic recovery (however feeble) now in its 6th year, how could the rate and number of people in poverty not be going down, and dramatically so? Instead, the number of people reported to be in poverty has increased since 2007 from about 37 million to 2014's 47 million, and the reported rate has increased over that period from about 12.5% to 2014's 14.8%.  On a longer time scale, the failure of poverty to yield to the extraordinary measures to end it is striking: a Census Bureau chart reproduced here shows that going all the way back to 1966 the official poverty rate started just above its most recent level and has never achieved any kind of downward breakthrough; the lowest recorded level was 11.1% in 1974.

The only way this is possible is that the government's "poverty rate" long since ceased to be a bona fide measure of physical deprivation, and instead became part of the ongoing efforts of the bureaucracy to grow the government in general and the anti-poverty programs in particular.  The bureaucrats well know that the best tool to advocate for more anti-poverty spending is a continued high poverty rate, while any dramatic decline in measured poverty would cause many to conclude that the poverty problem had been largely solved and the enormous funding could be cut.  And thus, as more and more anti-poverty programs and spending have been added to the system, definitions have been contrived to exclude the spending from the measure of "poverty," and surveys have been constructed to get answers that obviously omit large amounts of income and benefits received.

Meanwhile, politicians have shown little interest in devising a measure of poverty that meaningfully relates to physical deprivation.  Indeed, at the heights of cynicism, the Obama administration has backed an effort by a gaggle of bureaucratic gnomes to replace the official measure of poverty with a so-called "supplemental poverty measure" that is based on a standard of relative rather than absolute income.  In other words, the "supplemental measure" is explicitly designed to make defined "poverty" impervious to reduction no matter how much taxpayer money is spent to eliminate it.  Fortunately, the "supplemental measure" does not seem to have caught on much with the press or the public.  Meanwhile on the Republican side, the GW Bush administration did not pursue coming up with an honest "poverty" rate, nor is there much pressure to do so from right-side think tanks and pundits.  Perhaps Republicans just enjoy bashing the left with the obvious failure of their programs. 

But suppose we got a new President who decided to have the Census Bureau report honestly to the American people as to a real measure of actual poverty, that is, poverty in the sense that most people think of it, namely, physical deprivation.  By physical deprivation, I mean, clearly inadequate shelter, not enough to eat, insufficient clothes, or some combination of one or more of those three, plus lack of access to any government program to provide those things.  What could a new President do to move toward such a more honest measure of poverty?  As soon as you start to look into it, you realize that obvious modifications of the methodology in the direction of simple honesty could quickly reduce the reported rate of "poverty" by 90% or more.  For example:

• In an April 2, 2015 article in the New York Review of Books titled "The War on Poverty: Was It Lost?" Christopher Jencks of Harvard made what he called a "first approximation" of what would happen to the official poverty rate if just three adjustments were made: (1) counting in-kind benefits like housing and food assistance that currently are not counted, (2) counting the EITC, that currently is not counted, and (3) using the Commerce Department's Personal Consumption Expenditure (PCE) index, rather than the CPI, to measure inflation since 1964.  In round numbers, USDA's nutritional programs distribute about $110 billion per year; HUD's budget is about $50 billion per year (which way understates the aggregate value of housing assistance); and the EITC distributes about $61 billion per year.  Jencks concludes that counting housing and food assistance benefits would reduce the official poverty rate by 3.0%; that counting the EITC would reduce the rate by another 3.0%; and that making the inflation adjustment would reduce the rate by 3.7%.  Thus, making those three adjustments would reduce the poverty rate by almost 10%, to about 5%.  Jencks does not give details of his methodology, and his estimate for the reduction from counting the vast housing and nutrition programs seems ridiculously small to me.  But let's accept it as a first step.
• Next we have the fact that the poverty rate is calculated from a survey (the CPS) in which (according to Robert Doar in the Wall Street Journal on November 19) respondents are known to omit as much as half or more of the government benefits that they receive; and the Census Bureau, in tabulating the results, just takes the information provided by the respondents without any kind of follow up or double check.  Just counting unreported benefits received could easily knock another couple of points off the poverty rate.
• Dare I mention the underground economy?  Somehow the Census Bureau, Jencks, Doar and literally everybody else just leave it out when calculating a poverty rate.  Perhaps the most thorough scholarly treatment of the size of the underground economy is the Cebula/Feige study of 2011.  That study estimated the underground economy in the U.S. at $2 trillion per year, meaning that around 18% of all income goes unreported to the government.  Of course, not all of this money goes to those deemed to be in poverty, but clearly much of it does.  Please note that the total aggregate amount of income by which all people deemed to be in "poverty" by the Census Bureau fall short of the official poverty thresholds is only about $300 billion.  The underground economy exceeds this amount by a factor of almost seven, meaning that if only a seventh of the underground economy goes to those deemed "poor" there would be no poverty left -- and that's without considering any of the other adjustments discussed here. 
• And, believe it or not, there still are multiple big items that we haven't gotten to yet.  College scholarships and fellowships covering room and board are not counted in the current official measure of poverty.  Do you mean that living in the dorm and eating in the dining hall at Harvard counts as poverty?  They get away with this nonsense because in big towns like Boston the effect of counting well-off students as "poor" gets lost in the data; but then you look at a small affluent town like Ithaca, New York dominated by two colleges, and you find out that it has a reported official poverty rate exceeding 40%.  Ridiculous!  Another omitted category is loan proceeds, meaning, for example, that early retirees living off their home equity by means of a reverse mortgage are counted as "poor."  And finally there is what is likely the biggest category of all: support from family and relatives.  Yes, heiresses who take a year to travel the world on daddy's money are counted as "poor."  Hey, their income is zero.  They are living on less than a dollar a day!

I do not have an exact figure for where this would come out if all the adjustments are made.  But I can confidently assert that making appropriate adjustments to exclude from "poverty" those who are suffering no physical deprivation would reduce the reported "official rate" by 90% or more, that is, from close to 15% to more like 1.5% or less.  And, just as the Obama administration cooked up its "supplemental poverty measure" under a general Congressional authorization to study Census methodology, so could a new President us the same general authorization to put in place an honest measure of real poverty.