Feed on
Posts
Comments

Here a few of the interesting working papers in economics that were released today:

Evaluating Durable Public Good Provision using Housing Prices
by Stephen Coate

Recent empirical work in public finance uses the housing price response to public investments to assess the efficiency of local durable public good provision.  This paper investigates the theoretical  foundations for this technique.  In the context of a novel theoretical model developed to study the issue, it shows that there is little justification for the technique if citizens have
rational expectations concerning future investment in their communities.  An example in which investment is chosen by a budget-maximizing bureaucrat is developed to show why the technique can falsely predict under-provision.  The technique is valid, however, when citizens have adaptive expectations, believing that whatever provision level that currently prevails will be maintained
indefinitely.

Wintercow: There is of course a “big” debate among economists about the nature of people’s expectations. The result here is derived from the fact that current home prices reflect not only the characteristics of the home and the current community amenities, including quality of neighbors, quality of schools, environmental amenities, etc. but it surely includes expectations about what the community might look like in the future. Readers won’t be surprise by me being flummoxed that people could suggest anything other than rational expectations. One piece of “evidence” is to examine just what happens when cities try to rebuild communities and promote urban renewal projects. I’m finishing up a great history of Buffalo and that place has had a century of inner-city renewal. It ain’t renewed. So of course economists might simply be misunderstanding expectations and therefore predicting “under-provided” levels of  public goods in their models, but I am skeptical. We are all statists.

 

Are Government Spending Multipliers Greater During Periods of Slack?  Evidence from 20th Century Historical Data
by Michael T. Owyang, Valerie A. Ramey, Sarah Zubairy

A key question that has arisen during recent debates is whether government spending multipliers are larger during times when resources are idle.  This paper seeks to shed light on this question
by analyzing new quarterly historical data covering multiple large wars and depressions in the U.S. and Canada.  Using an extension of Ramey’s (2011) military news series and Jorda’s (2005) method for estimating impulse responses, we find no evidence that multipliers are greater during periods of high unemployment in the U.S. In every case, the estimated multipliers are below unity.  We do find some evidence of higher multipliers during periods of slack in Canada, with some multipliers above unity.

Wintercow: Though I suppose you’d think I’m pleased to read the above, note that Ramey is the major source of studies showing small or no multipliers in government spending, just as there are a host of economists whose work strangely produces only large multipliers. In any event, what is important about this study is that it is shining a light on time periods when, if multipliers exist, they ought to be the largest. When they say that in every case the multiplier is above unity, that basically says we are wasting resources. If we spend $1 on some government project, we end up getting less than $1 in return. Pure transfers, ignore the distortionary effects of taxation, would generate multipliers of 1. So, consider a government housing program. If you give people housing vouchers, or even cash,the multiplier would be 1. If you, instead, provide subsidies to landlords, all kinds of environmental grants for building low-income housing, etc. you get less than a dollar of value for each dollar spent. Nice. 

 

 

On Financing Retirement with an Aging Population
by Ellen R. McGrattan, Edward C. Prescott

A problem facing the United States is financing retirement consumption as its population ages.  Policy analysts increasingly advocate savings-for-retirement systems, but are concerned with
insufficient savings opportunities with limited government debt. This concern is unwarranted.  First, there is more productive capital than commonly assumed in macroeconomic modeling.  Second, if the policy reform subsumes the elimination of capital income taxes, then the value of business equity increases relative to the capital stock. Phasing in a switch from the current U.S. system to a
savings-for-retirement system without capital income taxes increases welfare of all current and future cohorts.

Wintercow: So this reform will never see the light of day. The simple observation is that you should save for your own retirement. Even “classical liberal” wintercow would support a program of “forced savings” if we scrapped the entire system we had today.

 

 

Subways, Strikes, and Slowdowns: The Impacts of Public Transit on Traffic Congestion
by Michael L. Anderson  –  #18757

Public transit accounts for only 1% of U.S. passenger miles traveled but nevertheless attracts strong public support.  Using a simple choice model, we predict that transit riders are likely to be
individuals who commute along routes with the most severe roadway delays.  These individuals’ choices thus have very high marginal impacts on congestion.  We test this prediction with data from a
sudden strike in 2003 by Los Angeles transit workers.  Estimating a regression discontinuity design, we find that average highway delay increases 47% when transit service ceases.  This effect is consistent with our model’s predictions and many times larger than earlier estimates, which have generally concluded that public transit provides minimal congestion relief.  We find that the net benefits of transit systems appear to be much larger than previously believed.

Wintercow: I think the second clause in the first sentence gives this away. I don’t know of strong public support. Public transportation gets a hell of a lot of support from the cadre of cronies who benefit from the construction contracts and union contracts to operate it. I am positive there is strong public support for safe, fast moving and convenient roads. Nonetheless this paper is arguing that the existence of public transportation does indeed provide congestion relief on roads contra existing thoughts on this in the field. My two cents given the literature on this is it depends on what kind of transport we are talking about. If only buses, then I wholeheartedly expect this result to persist. But not from trains and high speed rail. It is very well understood that the construction of trains in particular (and trollies and the like) ends up displacing otherwise convenient bus service. And long-run patterns likely emerge that alter traffic flow and the way people use public transport. It is not at all hard to think about how the institution of trains here not only does not help congestion, but after a long-time of running, how new patterns of transport emerge that would promote congestion of the train commuters suddenly had no way to travel by train. We’ll report more on this after I read the paper more slowly.

 

The Value of Climate Amenities: Evidence from US Migration Decisions

by Paramita Sinha, Maureen L. Cropper

We value climate amenities by estimating a discrete location choice model for households that changed metropolitan statistical areas (MSAs) between 1995 and 2000.  The utility of each MSA depends on location-specific amenities, earnings opportunities, housing costs, and the cost of moving to the MSA from the household’s 1995 location. We use the estimated trade-off between wages and climate amenities to value changes in mean winter and summer temperatures.  At median temperatures for 1970 to 2000, a 1°F increase in winter temperature is worth less than a 1° decrease in summer temperature; however, the reverse is true at winter temperatures below 25°F.  These results imply an average welfare loss of 2.7 percent of household income in 2020 to 2050 under the B1 (climate-friendly) scenario from the special report on emissions scenarios (Intergovernmental Panel on Climate Change 2000), although some cities in the Northeast and Midwest benefit.  Under the A2 (more extreme) scenario, households in 25 of 26 cities suffer an average welfare loss equal to 5 percent of income.

Wintercow: Read the paper for details. Their model assumes that the decision to work-retire-earn no wages is independent of each other, and estimates these models only for people who move with wages. They also admit that the people that move are systematically different than the people that stay, but do not make corrections for this, or specify how they might differ (they suggest one way is that they are more sensitive to wage changes than stayers). So they seem to be finding that for the cohort of movers from 1995 to 2000, that they actually warming temperatures more than they avoid cooling temperatures. The climate models I have seen seem to suggest that the way warming will occur is that nights will be warmer and winters will be warmer. I don’t see how we can say much about how moving responds to this by running regressions on average temperatures. Indeed see page 11 of their paper for strong evidence that they do not seem to think of this, much less control for it. A brief mention is made on page 17, but again no control for this is included. In other words, it seems to matter whether the average of, say, 70 degrees F comes from an 80F day and 60F night versus a temperature of 70 all day-night long. 

To derive their result for the “nice” scenario, they use the IPCC estimate the AVERAGE summer temperatures will increase by 3.3F and the same for the winter by 2050. They find that people in Northern cities will be happier by a little bit while people in Southern cities will be much less happy (e.g. the people in Dallas, where temps are supposed to increase by 6F, will experience a loss of satisfaction equivalent to 10% of their income, or about $5,000 per person per year). There are a whole host of difficulties for their estimation technique, including the selection of functional form to describe preferences, but more importantly to figure out what wages workers would have earned in cities they did not choose to move to. Further, they are using data on wages and not compensation, which to me makes the study hard to accept given that nonwage compensation differs vastly across states in part due to differences in laws regarding health insurance. They also admit that they are assuming the vast majority of people in the United States, those who do not move, are basically irrational (read the paper for more on this). And remember this observation from their data, “In contrast, only about half of the movers who lived in the Northeast or Midwest in 1995 remained in their region of origin. On net, householdsleft the Northeast and Midwest for the South and West, a pattern that began after the Second World War and is predicted to  continue at least through 2030.

There are some major issues again with interpretation. This is a cross-sectional study, meaning at best (if they do the estimation properly) they are telling us how much people value-disvalue climate amenities when they make a decision to move. However, this does not inform in any way how existing populations feel about changes in their existing climate. Here is a simple analogy. We can run wage regressions and find that people who move from New York to Alabama (“the South”) experience a 10% increase in earnings, all else equal. But from that you cannot say a thing about how existing New Yorkers would have their lives changed if “New York became Alabama.” Indeed we may not be able to know.

But more important, the paper doesn’t even suggest in the abstracts and summaries that people can and would move if climate disamenities got large enough, and only on page 18 and 19 do they alert readers that they explicitly do not allow people to move when climate changes when coming up with their estimates. Of course, no such caution is made in the abstract of the paper that their results are not very robust, that their estimation is not entirely realistic and that even doing this their results are a (perhaps major) overstatement of the costs of warming. In other words, these findings are static estimates. They finally say so in the second to last paragraph. If living in Dallas got 10% worse and living in Buffalo was a 2% improvement, then when climate changes people can actually be BETTER OFF than before climate change simply by moving. And in the year 2050, moving is likely to impose fewer costs – physical and emotional – on people than it did decades ago. Indeed, I think you can draw a consumer choice problem that demonstrates that ANY change in climate can possibly improve the well-being of individuals – more on that in a future post. Furthermore, a close reading of the paper will show you that the presented/preferred results are not very robust and are coming from specifications that suit the “needs” of much research these days to be “significant” – and this I remind readers is already coming after a process that we cannot see at home or in the lab where researchers prepare their work to be more amenable to find statistical significance.  In addition, there is not a simple to read table illustrating the list and ranking of influences on migration decisions. That would help the reader put these results in context. 

The paper ends with an obligatory gloomy observation: “We emphasize that these losses do not account for adaptation to climate change: the losses do not allow households to move in response to temperature changes. So they should be regarded as upper bounds to climate damages. Estimates for the United States of market-based damages associated with climate change have typically been in the range of 1 percent of gross domestic product for an increase in mean temperature of 2°C (National Research Council 2010). Our results suggest that the amenity value of climate could significantly increase estimates of climate damages, even for moderate temperature increases.”

 

Does Federal Financial Aid Affect College Enrollment? Evidence from Drug Offenders and the Higher Education Act of 1998
by Michael F. Lovenheim, Emily G. Owens

In 2001, amendments to the Higher Education Act made people convicted of drug offenses ineligible for federal financial aid for up to two years after their conviction.  Using rich data on educational
outcomes and drug charges in the NLSY 1997, we show that this law change had a large negative impact on the college attendance of students with drug convictions.  On average, the temporary ban on
federal financial aid increased the amount of time between high school graduation and college enrollment by about two years, and we also present suggestive evidence that affected students were less
likely to ever enroll in college.  Students living in urban areas and those whose mothers did not attend college appear to be the most affected by these amendments.  Importantly, we do not find that the
law deterred young people from committing drug felonies nor did it substantively change the probability that high school students with drug convictions graduated from high school.  We find no evidence of a change in college enrollment of students convicted of non-drug crimes, or of those charged by not convicted of drug offenses.  In contrast to much of the existing research, we conclude that, for this high-risk group of students, eligibility for federal financial aid strongly impacts college investment decisions.

Wintercow: yet another cost of the drug war. 

 

One Response to “Morning Morning Research Roundup”

  1. Harry says:

    That piece about multipliers is a good example of using language to bore and deceive. I guess there must be graduate courses in B******.

    First, “unity” is a $50 word for the more commonly understood “1”. And who thinks that 1 is a “multiplier” in any meaningful sense? Moreover, a number less than 1, which it ALWAYS is in physical and economic efficiency is more properly thought as a divisor.

    I cannot believe we pay people to write this stuff.

Leave a Reply to Harry