during its production. a november 10th "wall street journal" article cited an example of which a seat had parts from four u.s. states and four mexican locations. nafta makes the u.s. one of the most attractive manufacturing locations in the world because of value added productivity of both canada and mexico in one integrated north american supply chain. if we could complete free trade agreements with asia and europe, the u.s. could, in fact, become the undisputed champion in manufacturing once again. withdrawal from nafta would have massive repercussions. thousands of u.s. companies would have to ship their supply chains at great cost and disruption to their businesses. americans should understand that pulling out of nafta does not ensure that production in mexico would come back to the united states. in fact, it's possible that many u.s. manufacturers would either

find suppliers in other countries or use mexican production to export to other markets because mexico has 40 plus free trade agreements, double our level. >> you can watch the rest of this discussion online, c-span.org. we leave it here to go to a forum on public investment, due to wrap up 3:30 eastern time. and the earlier panels from this discussion available online, c-span.org. >> thank you very much. >> okay. let me get started as everybody is coming in to sit down and make sure, first, we need to really thank david and louise and carrie who put together a wonderful program and they have been good at nudging various people to do what they should on

time. so i want to just start out with some general comments and then turn to specifics. so, first off, education is an excellent investment. both for individuals and for society. i think we can -- i want to make this case in terms of an efficiency argument. okay. failure to close the gaps which are really big and growing in educational attainment is likely to both hinder economic growth, and also increase the burden on taxpayers over the long-term. and i think there are just two big takeaways from this. first, money matters enormously in education. but that's a necessary, not a sufficient condition. and then secondly, funding really needs to be matched to a commitment to accountability and in the education space innovation. and the federal government has a big role here. okay. that said, the title of this

event includes, i believe, from bridges to education. and it is worth noting why education is different and arguably a good bit more complicated than the challenge of building a bridge. okay. the economics of building a bridge is tough enough in terms of the allocative problem in terms of who pays it, how do we do it at the lowest cost. but there are blueprints, there is an engineering solution to building a bridge or building an airport. education is more difficult in that we're still trying to uncover the underlying technological process that is what we should do with different populations of students. we're learning a lot, but these are really hard problems. and again, it is worth making that distinction. secondly, another sort of

opening point is just about every big volume on education starts with some bit of a narrative about falling behind. it is certainly true that the u.s. lags other countries in terms of test score growth, and changes in college completion. but it is also worth putting a more optimistic note on the table, that is over the last couple of decades, we have actually seen some growth in test scores and in the early grades, they have been stagnant in the middle grades. we have seen some reductions in high school dropout rates. in the most recent decade, we have seen increases in college enrollment rates. we also have seen increases in college completion rates. that said, those gains are not spread evenly across the population. and we should be increasingly concerned about the degree of inequality. what is more is i think we're going to talk as we go along in this, we know a great deal more about what works. and actually what doesn't than

we did two decades ago. i'm sorry, russ whitehurst isn't here. some of this owes a big debt to the founding of the institute of education sciences in 2002. but there is a hopeful note here, we made some progress, there is a lot more to do. okay. two messages we want to use evidence to shape policy. we also need to innovate in this space. my assignment, what are high return investments in priorities for federal education policy? i decided to make that a little bit narrower. there are two book ends here that are really important. pre-k, talk to diane, she'll answer any questions about pre-k. and also i think sometimes forgotten piece of graduate education and funding the sciences. but we have got to set some boundaries here. and indeed i'm going to set even a better boundary here, i'm going to do something that economists are pretty good at.

dick and i have colluded on this enterprise. we are going to exercise some division of labor here. i'm going to concentrate on the post secondary margin, really my comparative advantage, he's going to concentrate on k-12. that said, we have a lot to discuss in this sphere. i think within k-12, you can think about three buckets, score accountability, school choice mechanisms, teacher preparation. we're going to hit on those i think through the course of this 40 minutes. i'm going to talk in terms of the post secondary sphere, in terms of student financial aid, college choice, and then the supply side of higher education. i want to say just a quick moment about the federal role and education policy here because it deserves some comment. so overall we spend about $1.2 trillion on education each year.

the federal role is about a quarter of that, okay. and in that quarter, that's about $100 million or equally divided between roughly elementary secondary and k-12 and another $100 million or so that is off budget in terms of the student financial -- in terms of student loans. you might ask, what are the big missions of the federal government and in the education sphere? i'm going to hit three here. the first is addressing credit constraints, credit constraints are one of those market failures that economists love to teach about in introductory classes and really matter in education. there is good reason to think that individuals cannot fully finance worthwhile investments in education. there is a compelling government role in this area. second, what i'm going to call limited but certainly not zero role of the federal government in regulation or auditing the

use of its funds, preventing the worst outcomes. third, and this is where i think hopefully we'll have more discussion is that the federal government really has an advantage in funding research, innovation and development. these are innovations that when we discover something that works in one area, we can spread them around to other communities. many of these things need to be done at scale. and it is only -- and the federal government is actually well positioned, not necessarily to execute these experiments, but to at least cede them. this -- the next preliminary i want to put on the table is just a reminder here of really the degree of inequality, that challenge that we're facing. in educational attainment. as kristen noted, these things start early.

the gaps start very early before kids enroll in school. but continue on. here i got the entrenched elementary secondary achievement gaps, the blue bars are comparisons of the quinn tiles by education, the red bars reflect the black/white difference. these are very -- i'll skip the precise characterization, should think about these in terms of grade levels of achievement, and -- or they are very meaningful in terms of grade levels of achievement, and what is striking here is how large the economic gap is actually relative to the race gap. so if you look back 50 years ago, the race gap would have been larger than the economic gap. but things get worse as we go on to college enrollment. the left-hand panel is enrollment by family income, the lowest quartile on the bottom, the highest quartile to the

right, the right-hand panel is enrollment, the left-hand panel is completion, the dark dashed line is essentially the behavior of those students who were making college going choices in the '80s. the lighter blue line is students making college choices at the beginning of the 21st century. and the takeaway is from these, even if you're in the back, are, first off, there is a positive gradient. second, the gaps have widened and they widen really quite markedly over time. so we have got about a 30 percentage point difference in enrollment rates, not much more you're going to push up that enrollment rate in the top quarter quartile. and an increase at the bottom, about the increase at the top of this distribution between the top and the bottom.

now, one might rightly note these are not adjusted for differences in achievement on entry. again, if you do this additional calculation, you still see very large gaps and they have increased over time. so on the order of about 16 percentage points in college completion. okay, that's the preliminary. that's the problem we need to address. as i say, i'm cede something territory that will come back to on k-12, but i want to talk about higher education and higher education may actually get some billing here in the d.c. policy environment, given that the elementary and secondary act was just reauthorized, what is on the table is reauthorization of the higher education act, which somehow or another congress hasn't quite gotten to yet. and within this rubric are the big programs in title 4 financial aid programs which you probably know as stafford or direct student loans and pell

grants. but the other sort of key topics that i want to touch on are the college choice problem and essentially the supply side problem. okay. federal aid. grant aid. we have two slides on student aid here. again, i think we have really compelling evidence at this juncture from sue, from work i've done on the gi bill, the transparent grant aid can have a very positive effect on collegiate attainment. we have actually grant aid on the table now, or effectively grant aid in the form of the pell grant, which we spend about 30.6 billion per year on. and also implicitly the tuition tax credits. which amount to another 18.2 billion if you look back in 20 10 at the peak enrollment period post great recession, those numbers total to about $60 billion. the bad news is that even though

these programs have two features that one might really like in a student aid program, that is they're portable, so the -- like vouchers, secondly, they're means tested. problem is they're not transparent. and because they're not very transparent, particularly the tuition tax credits, they're not having necessarily the impact that we with like to see on student enrollment, and more importantly helping students to finance really worthwhile collegiate investments. and i think i'll draw your attention to two issues here. first, the tuition tax credits. separate from the problem of really a total lack of salients. they don't matter to you if your parents don't get paid another 18 months from when you apply to college or you're not going to see that in terms of your tax

benefit. and, again, at that point you may have lost interest or -- it is not going to affect your decision and indeed that comes through very clearly in the research literature. many students don't even know about them. the pell grant is actually a bit of a challenge here, because it serves such a broad umbrella of students that it is not very well targeted. i want to draw your attention -- so the pell grant generosity has actually increased a bit in the last decade. i want to draw your attention to the column on the right here, the proportion of students who are independent, that is likely over the age of 24, they have young children of their own. now, one of the challenges is designing an age system that meets the needs of this population, who are likely responding to near term economic shocks as well as the needs of

students who are recent high school graduates. and the system of needs analysis that we have doesn't accomplish either objective very well. and so given i'm running out on time, we'll come back to this, but there are excellent recommendations by a brookings, i believe, supported panel called rethinking pell grants that would serve to essentially divide the resources with separate needs analysis systems between pell grant adult program and pell grant young program. loans. everybody's favorite question here. i'm going to just simply note that there is nobody -- anybody who read a newspaper in the last ten years knows that there has been much attention with headlines like a generation hobbled by the soaring cost of college. contrary to what some newspapers would have you believe, the number of undergraduate students

who are drowning in six figures of debt is more like one in 30 rather than the median or the mode. so there is a point of just getting the numbers right that are important. also there is an important study i believe presented last year by adam looney and constantine yanelis that looks at what is a real increase in default rates, which occurred over about the last eight years to increasing to about 10.1% from 5%. and they ask why. and there are two big factors that are at play here. the first has to do with the changing student population, a shift that is the students who are most likely to struggle, are these older if you will nontraditional borrowers, and second, a shift in the type of institutions that the students are attending. so when you take out the compositional changes, the loan

issue is different than it has often been characterized. so this brings us to the question, there are real questions as to whether some of the students struggling are really being buried because they may be enrolled in a college that had weak returns in expectation at the worst example of this would be those institutions that have turned out to be downright fraudulent. and there is a question of can we help those students avoid those choices, which is a college choice problem, and a question of what do we do with students struggling in repayment, given i think i'm nearly at negative time here, i am just going to hit one point here on my list, which is one of the most popular policies from both sides of this aisle has been discussion of income-based repayment. and there has unfortunately been little attention to how this

program actually affects the liabilities of the federal government. recent gao report notes that the liability is now increased to about $74 billion, which is about triple what it was estimated to be. essentially what you're doing is you're trading an insurance mechanism for more moral hazard and adverse selection. the primary beneficiaries are going to not be those who have borrowed a little bit and really are struggling with small amounts of debt, but turn out to be those who are getting -- forgiveness for graduate borrowing. i think i'll come back to this. can i have two minutes and then i'll -- two. okay. we'll come back to this. college choice. there is much to do here. this is a case where we need r&d sponsored by the federal government. it is a real -- again, there are

two groups of students who are not very well guided at this point, particularly the older nontraditional students who don't have access to either purists who are going to college or traditional guidance mechanisms. i think they're very interesting experiments we can do there. supply side. again, resources matter enormously, okay. and if you see what is going on, public institution resources per student have declined markedly. these are constant dollar. and the issue here is how can we -- how can we encourage greater state funding, we see increased straight fiction. this is an issue to address. but really the challenge -- federal government to support productivity enhancing innovations. again, are there consolidations

that can be supported. and literally the billion dollar question is can technology change education productivity in the higher end space? okay. main takeaway, again, we have got -- accountability, addressing market failures. i want to end on this final note, which is we're doing better. we're learning a lot about how markets work in education. but there is room for more investment here. and just as a relative point, this spending on research, on education, is about 279 million a year, which is about 102 times -- or spending on nih is about 102 times greater as is spending on nasa, about 43 times

greater. there is room for more investment here and i think many high return projects to think about. let me turn this over to prof s professor renee who will take on the k-12 side, i believe. [ applause ] >> sarah has written a thoughtful valuable paper. i agree with the role, she is for federal education policy. as she mentioned, we agreed i would focus on recent research, how recent research informs the design of federal activities in three areas. accountability, teacher policy and school choice. just a few words on context. inequality in educational outcomes among the 50 states, each of which has its own educational system, is very high. this is evident in the results of national assessment

educational progress. low quality schools are associated with low rates of intergenerational mobility. this is worrisome because the promise of upward mobility provides a lot of the glue that has held our pluralistic democracy together. so improving education, especially in states where low quality state systems and especially children from low income families should be a goal of federal education policies. so what are the policy tools? as sarah writes in her paper, funding and regulation are the primary sets of tools. the federal government has attached rings to aid and this has affected the actions of states and school districts. good research on that. as reaction in some states to the obama administration's tying great recession -- post great recession education funding to adoption of the common core has

shown, regulations are not very popular. such incentives to alter behavior. the every student success act that moves the design of accountability systems firmly to the states is still a regulatory role and still very much up for grabs. the details of these federal regulations will look like. i'll come back to that. turning to first of buckets, accountability. very important as sarah emphasizes. it is also very difficult to get accountability right. a litmus test of that would be whether the accountability system encourages skills a teacher is to work in high poverty schools. that's a test that most accountability systems will fail. now, what are our federal roles in this regard. i see four. one is the auditing function.

strong support for national assessment educational progress is absolutely critical. also, opportunity would be cautionary to states to participate in international assessment programs such as pisa and times. three american states participated in the 2012 assessment system. all these dates have by reports -- percent proficiency on their own state tests show they're doing about the same. but in fact, on the pisa test, one of those states had average scores way above the average, one had scores way below the average. again, that's the importance of this auditing function and helping states to benchmark how they're actually doing. so that's the second. second, signal openness in -- to

innovation and accountability systems so it is a spark innovation in school design. currently all state accountability systems are based primarily on student math and reading scores. now, the skill are clearly important and matter. but there are at least four well done research studies showing long-term effects of interventions designed to improve the lives of low income children that did not affect test scores. one is moving opportunity. movement took place before the age of 13, no effects on test scores. that's that suggests the importance of thinking upon accountability in a broader sense, the greater availability of data on college, on crime, on labor market participation wages suggests the possibility of designing much more creative accountability systems. i think encouragement of that

would be valuable. particularly encouragement of innovation in the design of education for our teenagers. if you look at the scores, you see -- sarah mentioned, improvement in scores of 9-year-olds, no improvement in scores of 13 to 17-year-olds and very large gaps by race ethnicity and income. we really need to have -- find new ways of designing education for teenagers, and some of those ways to try and include more connections to the world of work. so i think the feds can signal an openness to accountability systems that would support innovation in these areas. samples of things tried with some success, small high schools in choice new york city, early college high schools, national guard challenge, some urban charter schools. third will be support collaborations of states so work on the design if new educational

options for teenagers. it would be great particularly if states with weak systems collaborate with those whose stronger systems want to fund them. finally, of course, fund research on the consequences of innovative accountability. we would hope to see significant variation among the states. on teacher policy, you see all over the place, the research that shows what every parent knows, teachers matter and they're a big variation in teacher quality. what is much less attention, however, is very good research findings showing the performances of novice teachers and the rate at which they improve their performance depends on the skills of their grade level colleagues and on the -- and the quality of the environment allowing them as adults to learn. so that's critically important and you think about accountability systems, where a novice teacher is placed will have a big impact on how well she figures into accountability.

now, don't unfortunately know very much about how to design systems that provide both accountability, the combination of accountability and support. so that's an area where we really need to do more research. of course this is with school district central offices are supposed to do, provide this combination. very few know how to do that. so i think there is real need for research in that area. another area related to this and which i think research could be promising is looking at how charter management organizations are trying to design that same combination of supports and accountability for the school in their network. some initial results are somewhat promising, but, again, has not been a systematic research program. school choice. clearly potentially valuable widespread support. but the thing that has not been talked about that is critically important, might be called peer

group influences. i want to quote the brief quote from a recent -- research working paper, exposure to a disruptive peer in classes of 25 during elementary school reduces earnings at a 26 by 3 to 4%. we estimate that the differential exposure to children linked to domestic violence explains 5% to 6% of the rich/poor earnings gap in our data. so you can clearly understand any system of competition, competing for students, what kind of students do you want to avoid? students who are likely to have those kind of deep seeded behavioral problems that come from domestic violence at home, perhaps domestic violence that their families observed in central america before they came to the united states. that's, i think, does not mean the choice is not a good idea. but it means enormous attention needs to be played to where the

children go to school and the consequences for them and for the children who are in school with them. now, laws and regulations governing charter schools vary enormously from state to state. there is very little systematic knowledge about the extent to which these regulations affect which children go where and how that affects outcomes. that's a very promising area for research. on vouchers, another area of school choice. there are lessons i think from observing chile, the country had national k-12 educational vouchers since 1981. up through 2007, the value of a voucher did not depend on the family's income. nor was there very much accountability for private schools. it has been well studied and the may main consequences of this are

t two. in 2008, chili dramatically changed the educational voucher system. the vouchers for poor kids are worth 50% more than those from affluent families. there is a school receives a concentration bonus if it -- a percentage of poor kids and certain percentage accepts vouchers. that led to improvements in math and reading achievements and closing the gap between low and high income kids. people ask what do you think of vouchers? the only sensible answer is i have a few questions of how it works, the details matter enormously. be sure that the rules governing

recessi regulations don't hinder innovation, particularly in design of education for teenagers. support research, especially in the consequences of state and local initiatives in in these as of accountability, teacher policy and school's schois. and all of these areas there's a great deal to be learned and about the consequences of details of these policies for the distribution of student achievement. thank you. [ applause ]. >> thank you so much the

presentations were both so interesting. talk about putting a lot of material in to just a few minutes. it's really great. i have so many questions, we won't get to most probably. let me step back and just sort of ask a broad question which is -- what's the major problem to be addressed in education right now? do you think it's like the whole system isn't doing a good job of educating children? or is it really, you know, it's doing pretty well for most kids but really poorly at the bottom, kind of just broadly. both of you. >> well, why don't i talk about the kids. again, there's not one system. there are 50 systems and they look fundamentally different, and they have very different outcomes. and those outcomes matter as has been shown. the federal government spends less than 10% of the money. it has some leave with regulations that are not very popular. so the problem is there isn't one system. >> i think in higher education you're talking about a diversity

of over 4,000 institutions, and they are very, very different in terms of the students they serve, their focal mission. i do think that in terms of picking one issue, it is the success of low and actual moderate income students and their capacity to both make good college choices, to finance those college choices and then ultimately to complete. >> and staying on this kind of big picture, some people would say, look, you noted in your paper that real per capita spending in the k through 12 area as almost doubled in 30 years, but you say the results have been modest. so, someone says why are we going to spend more? every time we spend more we don't do very well. maybe you aren't advocating spending more primarily, so one, are you advocating spending more on education? two, how do you respond to that criticism that like more money doesn't seem to matter?

>> a couple things. first of all, there's been a new work. there's a very nice paper economics where they show the impact of funding on student outcome is greater than previously thought. diane and her decolleagues have second paper that shows the same thing. i think the key piece of it over time while the u.s. has not gotten great on accountable, it's better than in 1975 when the education act was first passed. doesn't mean that all money is used well, by any means, but i think we're beyond the point to say money doesn't matter because with title 1, again, it goes to 14,000 school districts. that doesn't make any sense at all if you're thinking about having an impact on the lives of poor children. >> that it goes so broadly you're saying? >> that's a political reality

that doesn't make sense. >> i think in higher education, you know, again, averages are deceiving here. and there's no question that money matters. it depends on who you are as to whether resources have increased or not. if you're a student at one of the most elite universities in the country, resources per student have increased. if you're a student who is attending a community college or open access foyer institution, it's likely that given reductions in state funding that resources per student have decreased. so, again, there's a lot of head jer nayty. i think we've come around to see that on whole resources really do matter, at the same time, you know, that's a necessary but not a sufficient condition for education success. and there is room both in k-12 and in higher education for innovations that essentially

increase productivity. that is, improve student learning without changing the costs. >> so, talk a little bit about school choice, start with the k through 12. i have a factual question. the every student succeed act has already been enacted but a lot of resgs have not been promulgated yet. what is the scope for the next administration? how much lee way do they have to change education and in particular to think about charters and vouchers and what are the facts on that? >> i think a lot will depend on how the president-elect uses the bully pulpit, will be one thing. but again, you know -- i think

the the effects through -- well, for example, there are these very detailed issues. there's language about the allocation of spending across schools in the same district and there's a big fight about how do you count teacher salaries. salaries in schools that serve primarily middle class kids are higher because teachers are much more experienced. well, there's a big fight about that. i think the -- in the area of research, i don't think they have much leeway over charter schools, except perhaps to encourage more attention to these regulations through research. the fact -- you know, we always hear about the cradle study that on average charter schools aren't any better, but to my knowledge, there's not been systematic look at these by states. and the regulations on charter

schools vary enormously in terms of when they serve and what outcomes they must show and whether if they do a poor job whether they're eliminated or not. >> so let me just jump in here. i think we're on the same page entirely in that this notion that we have -- we have 50 state experiments going on and then within that about 14,000 school districts. and to -- we have an increasing body of evidence on matters like school -- school choice, the charter programs as well as vouchers, it's good evidence but it's honestly a little bit mixed in various forms. and it is not the kind of evidence that i think certainly -- i don't think dick or i would be comfortable in suggesting that any piece of it is so definitive that it should suggest a specific set of federal regs on charters, vouchers or teacher compensation

for that matter. we are learning a lot. you know, there's room to learn more. it's imperative to collect data, to assess it carefully, but that evidence actually doesn't, i think, support strong federal policies in this area, beyond this important what i'm going to call an audit function and also this function of making sure that there's really the worst kinds of fraud and poor performance don't persist at the bottom of the distribution. >> one other comment. after the esea was passed in 1965 that provided title 1, the first significant federal funding for compensatory education, the department of education with the support of the administration used the with

holding of title 1 funds as the stick to get a southern school district to comply with the civil rights act and effectively desegregate schools. now, it was not popular at all, of course, but it did have -- it did achieve its objective. again, that's the question of whether federal government is willing to use its regulatory power because 10% is not a lot -- big percentage, but on the margin it's significant dollars, but this does take a pretty heavy hand. >> okay. i have more questions. let's move on so we don't have time. let me ask about teachers, which i thought there was a lot of interesting stuff in your papers and in your discussion about teachers. within this idea within three years of teacher's kree, you basically know if they're a fabulous teacher or horrible teacher and you don't know where they are in the middle. the tails show up. one thing i was wondering you did mention that if teachers are basically -- they are people who

will be great teachers and people who will be horrible teachers, then some of these pay for performance or schemes like how could they even be expected to have been impact? how much of it that teachers aren't trying hard enough or don't have enough incentive to try or there are some people who are wonderful at this and other people who should not be teachers? now, you mentioned that there's evidence that within the first three years you actually can maybe turn, create good teachers? >> well, couple things. first of all, the newest work shows that in the right setting, with the right support, teachers improve well beyond the first couple of years. new work is quite good, but only in those settings. now, i think this pay for performance, that's the current name. the old name was merit pay. mathimatica has done a $13 million study funded by ies on the effects of variety teacher

performance plans. my interpretation of that evidence is almost no effect, except in what in most cases most teachers got extra money. in those cases, of course, they were quite popular. not a very powerful strategy, i think, for -- but i would distinguish that from the situation of providing extra money to work in difficult situations. combat pay will not do the job if there's not support to do it -- to actually do the job, but if the job has particularly a longer school day, a longer school year and it's quite demanding some extra pay for that can make sense in subject areas you're having difficulty, extra money for that, but that's very different than the performance-based pay. >> let me ask you about some things that you didn't mention. so one of the things that people normally think about, you know, when you're trying to buy a house and choose where to send your kids you look at class size. and i know there was some

controversy over whether or not class size matters. that's clearly something where you could spend money on. what is the current thinking on the importance of class size to achievement? >> well, i think the best -- you know, the analysis of the tennessee star experiment shows that having smaller class sizes in kindergarten makes substantial difference, particularly in schools serving higher concentrations of lower income kids of color. as you get to higher grades, the research is much -- is not nearly as clear cut. >> i think there's also this question of markets and implementation matter enormously in this many. the tennessee class size experiment is very different than the rollout of reduced class size in california, which again the incentives were to reduce class size effectively independent of other educational considerations. so actually kids probably a little better off in a slightly

larger class than a class that combines across grade level. again, you want to be careful in terms of how you do these rollouts so that you don't end up having the more affluent districts effectively buying the very experienced teachers from the low income districts. so again, design matters, implementation matters enormously in how these policies are put into play. >> yeah. let's move on to higher ed quickly and i'll get to some audience questions. so, you talk a little bit about school choice and that's a big problem. people are not going to the right schools. is that like really low-hanging fruit? is that something that would be very difficult to change or something that you would be not that difficult and could really have a big effect? >> well, it's something that can really have a big effect. you know, it is very data dependent to do it well. it also is very differentiated,

so done correctly you want students informed in ways that are very personalized, that take into consideration their geography, their achievement and their local market options. and again, i think it's very important for students to understand net price so the difference between the posted tuition and financial aid, similarly students need to understand meaningful differences about how effective different colleges and universities are. you're not going to find a student who says they don't want to go to a very good college. all students say they want to go to a really good college. the problem is that students often can't distinguish between institutions based on their graduation rates or their resources per student. >> do you want to -- >> not on that. i want to come back to the class side if we could. >> last question for me is going to be about -- one of the things

you mentioned again looking for things that are easy and could have a big bang for the buck is closing some of the really terrible places. is there a federal rule both in closing and in sort of supporting states and localities in making sure there's enough community college seats available for these students to go to so you don't end up having people not have a place to go at all? >> so again, i think that there is -- you know, the human cost of institutions where my favorite -- there are institutions out there where the on-time completion rate is actually less than the default rate. that should be probably a clear indication that an institution isn't functioning as intended and is probably not using title 4 aid well. i think it is imperative to not let these institutions go on too long in this situation. the accreditation mechanism is

nominally supposed to address this. i think it is wholly ineffective and probably wasteful in the administrative time it takes. it's a lot of paperwork burden and it's not properly identifying the poorly performing institutions. >> questions? why don't you go to that lady there, that woman there. >> hi. i've done work for the department of education. i'm curious what you think of some of the work that ies has spencered, has that really made a difference? has that impacted the choices? the role of the regional education labs? where is that going and what has been its success rate? >> i think sara and i both feel that ies has -- compared to what came before -- has contributed to a marked improvement in educational research.

and it's not while ies does fund a great many randomized control trials, that's not all it does, contrary to what some people might think. i think the labs are mixed bag. i think they for a long time they had their own lobbying agency they put a fair amount of money put aside just for them i'm not so sure that's the best way -- in fact, i think it's not the best way to use scarce dollars. i think more competition for funds makes more sense. >> again, i'm in the same view about it. i would emphasize on the i es grand funding. at its best the ies is investing in a portfolio of projects and they wouldn't actually be taking enough risk if all of them had

big positive or had -- all showed positive effects. you know, part of this is to actually take good ideas that are theoretically driven, look at the data, come up with a good way to assess whether something works, what it's costs are relative to benefits. i'm actually less certain that it's had a big impact on practice, but i think it has forced some discipline on researcher activity. i want to come back to a notion that you made in your remarks, but again, this idea that collaborations may be really high return, both among states, among districts where you're going to get an economy of scale that you can't get in innovation and system development if you

expect every small metropolitan area to say develop a teacher performance review program independently. there have got to be enough similarities between des moines and topeka where there are gains from collaboration there and that's worth funding. >> last question. >> keying off of dr. summer's chipping paint summer this morning, environment is a very important thing to kids being motivated or willing to learn, and it's also a good segue to helping them go into the junior high, high school years. so it seems to me there's an opportunity to both -- well to try and couple some policies with some local efforts to improve physical infrastructure, physical environment of the schools as well as some of the social environment things as well. i think a lot of research has shown that social environment has a big impact on student's success.

so, are there considerations or some real policy ideas that can be brought forth from that? >> sure. i mean, having schools -- kids and the adults who work on them want to be in seems like something we as a country can surely afford and should do, and we haven't done them. many of our schools in boston are more than 100 years old. >> okay. i think that we better move on. i'm sorry there wasn't time for more questions, but thank you so much. this was such an informative panel. appreciate it. [ applause ]