The appointment of new federal chief information officer Steven VanRoekel comes at a challenging time for President Barack Obama. The national economy continues to be weak. Congress plans to cut trillions from the federal budget. And in the time leading up to the 2012 election, American voters remain cynical about the ability of the government to address important policy problems in an effective manner.

In an era of deficit reduction and public cynicism, the tasks facing federal officials are to determine how to do more with less and persuade voters the government can become smarter and more effective. There are going to be fewer dollars for virtually every federal program so it is important to figure how ways to innovate and perform more efficiently.

Former CIO Vivek Kundra sought to do this through encouraging agencies to move software applications to the cloud, consolidating federal data centers, improving transparency, and improving the information technology procurement process. It is important to continue this progress even as agencies are forced to downsize their operations.

As shown in the private sector, government administrators should use technology to cut costs, improve worker productivity, and streamline operations. This is not just a matter of using technology in more innovative ways, but changing the operations and culture of the public sector. Public officials must improve its data mining activities to identify fraud and abuse in Medicare, Medicaid, the Defense Department, and other domestic programs.

New software gives managers better tools to evaluate how money is being spent and whether it is fulfilling intended goals. If it is not, programs need to be modified or eliminated. The most important weapon in Mr. VanRoekel’s arsenal may be the scalpel as he goes through the federal government’s $80 billion IT budget.

Donald Trump has made waves during this year’s election cycle by taking a hard line on illegal immigration. This, however, builds on years of heated debate among policymakers. It is also an enduring hot-button issue in Arizona, which has passed several immigration laws over the years.  In 2010, the passage of SB 1070 brought national attention to this debate.  Deemed the strictest immigration law to date, SB 1070 sought to achieve “attrition [of illegal immigrants] through enforcement” by requiring law enforcement to detain any person whom they believed to be residing in the country illegally. Although SB 1070’s effects on individuals and families have been well documented, little is known about its impact on students and schools. To this end, we sought to estimate the relationship between the passage of SB 1070 and school-level student achievement.

We anticipated that anti-immigration policies would primarily affect children from the families of undocumented immigrants. Such effects could be observed in different ways. For instance, the emotional and psychological distress of these children could result in a decline in average test scores at the school-level. On the other hand, students might have left the country or the state under the threat of being deported in which case school-level test scores would rise (since these students often perform below their peers). To this end, we considered three scenarios: 

1. Immigrant children remain in the state but experience higher levels of stress.  As a result, average school-level test scores will drop while Hispanic enrollment remains the same.
2. Children of undocumented immigrants leave the state, which results in a drop in Hispanic enrollment accompanied by an increase in school-level test scores.
3. Or, the first two scenarios occur simultaneously and we do not observe any change in test scores since the two effects would cancel each other, but note a slight decrease in Hispanic enrollment.

In order to see which of these hypothetical scenarios is supported by the data, we first estimated the relationship between the passage of SB 1070 and average school-level reading test scores. We then attempted to unpack the mechanism through which such an effect might have taken place. To this end, we used publicly available data on school-level achievement and enrollment collected by the Arizona Department of Education (ADE). Given the targeted nature of the policy and the demographics of immigrants in Arizona, the majority of whom are of Hispanic or Mexican descent, we focused on schools that traditionally enroll large proportions of Hispanic students. We identified schools with high (more than 75 percent) shares of Hispanic students as those whose average achievement and student composition are most likely to be affected by immigration reform. We contrasted changes in school-level achievement and enrollment in those schools with schools that enroll less than 25 percent Hispanic students, as these schools are less likely to experience any changes as a result of tightening immigration laws.

Figures 1 and 2 show trends in the average percentage of students passing the state reading test and average Hispanic enrollment at these schools between 2006-2007 and 2011-2012.           

Figure 1. Average Percent of Students Passing AIMS Reading

Figure 2. Average Hispanic Student Enrollment

Clearly, the rate of growth in school-level reading scores was much higher for high Hispanic schools after the passage of SB 1070 in 2010 (Figure 1). At the same time, there was a significant decrease in Hispanic enrollment in these schools (Figure 2). Thus, it appears the second scenario is likely driving the patterns we observe.

The data also suggest that the trends for high Hispanic and low Hispanic schools started diverging before the passage of SB 1070 - after the 2007-2008 school year.  This happens to be the year that Arizona passed an even more restrictive, though less controversial, immigration law – the Legal Arizona Workers Act (LAWA). LAWA required Arizona business owners to verify the legal status of their employees using E-Verify, an online tool managed by the federal government. Although LAWA used a different mechanism, similar to SB 1070 it sought to achieve the attrition of undocumented immigrants from the state. 

We then would anticipate both laws to have similar effects on school-level achievement and Hispanic enrollment. Indeed, we estimated that LAWA likely led to an average increase of roughly 4 percent of students passing the state reading test at high Hispanic schools. This was accompanied by an average loss of 38 Hispanic students per school. Because the passage of SB 1070 was preceded by the passage LAWA as well as a language policy that would have affected treatment schools, disentangling the effects of these two policies is not straightforward. However, based on our analysis, we estimate that SB 1070 is associated with an average increase of between 1.5 percent and 4.5 percent of students passing the state reading test at the school-level accompanied by an average loss of between 14 and 40 Hispanic students. 

Despite the fact that we cannot pin down the exact magnitude of SB 1070’s effect on school-level achievement, our analysis shows that when Arizona passed restrictive immigration laws in 2008 and 2010, it looked as if the state’s lowest performing schools were improving rapidly. This, however, likely had more to do with the changing composition of schools as an indirect though anticipated effect of immigration policies than with policies aimed at improving student achievement. 

Despite this, the Arizona Department of Education took credit for these gains. Similarly, Arizona was recently recognized as one of the nation’s leaders in growth on the National Assessment of Educational Progress (NAEP) over the last ten years. Although wrongly attributing these gains may seem harmless at first glance, it is important to remember that Arizona is viewed by many as a model for controversial education reforms like school choice and high-stakes accountability. It is easy to imagine how policymakers might look at increasing test scores in Arizona and wrongly attribute them to these kinds of reforms. That’s not to say that these policies don’t have merit. However, if other states adopt education policy reforms under the assumption that they worked in Arizona, then they might find that these policies fail to deliver.

June’s jobs gains, released this morning, show that 287,000 new jobs were added in June, an impressive rebound after only 11,000 new jobs were added in May (revised down from from 38,000 at the time of the release).

This year’s monthly job gains and losses can indicate how the economy is doing once they are corrected to account for the pattern we already expect in a process called seasonal adjustment. The approach for this seasonal adjustment that is presently used by the Bureau of Labor Statistics (BLS) puts very heavy weight on the current and last two years of data in assessing what are the typical patterns for each month.

In my paper “Unseasonal Seasonals?” I argue that a longer window should be used to estimate seasonal effects. I found that using a different seasonal filter, known as the 3x9 filter, produces better results and more accurate forecasts by emphasizing more years of data. The 3x9 filter spreads weight over the most recent six years in estimating seasonal patterns, which makes them more stable over time than in the current BLS seasonal adjustment method.

I calculate the month-over-month change in total nonfarm payrolls, seasonally adjusted by the 3x9 filter, for the most recent month. The corresponding data as published by the BLS are shown for comparison purposes. According to the alternative seasonal adjustment, the economy added 286,000 jobs in June (column Wright SA), almost identical to the official BLS total of 287,000 (column BLS Official).

Data updates released today for prior months also reveal some differences between my figure and the official jobs gains from prior months. The official BLS numbers for May were revised down from 38,000 new jobs to a dismal 11,000. My alternative adjustment shows that the economy actually lost 6,000 jobs in May, down from 17,000 jobs gained at the time of the release. [i] The discrepancies between the two series are explained in my paper.

In addition to seasonal effects, abnormal weather can also affect month-to-month fluctuations in job growth. In my paper “Weather-Adjusting Economic Data” I and my coauthor Michael Boldin implement a statistical methodology for adjusting employment data for the effects of deviations in weather from seasonal norms. This is distinct from seasonal adjustment, which only controls for the normal variation in weather across the year. We use several indicators of weather, including temperature and snowfall.

We calculate that weather in June brought up the total by 25,000 jobs (column Weather Effect), but this should be considered a transient effect. Our weather-adjusted total, therefore, is 262,000 jobs added for June (column Boldin-Wright SWA). This is not surprising, given that weather in June was in line with seasonal norms.

It’s good to see the jobs numbers rebounding this month. The May number was somewhat affected by the Verizon strike. Also, it is important to remember that pure sampling error in any one month’s data is large, and that could explain part of the weak employment report for May. Averaging over the last three months, employment is expanding by about 150,000 jobs per month—a healthy pace, although a bit of a step down from last year.

a. Applies a longer window estimate of seasonal effects (see Wright 2013). The June 2015 to May 2016 values in this column have been corrected to remove a coding error that affected the previously reported values.

b. Includes seasonal and weather adjustments, where seasonal adjustments are estimated using the BLS window specifications (see Boldin & Wright 2015). The incremental weather effect in the last column is the BLS official number less the SWA number.

[i] Note that, due to a small coding error, my alternative seasonal adjustment for May, at the time of the release, should have been 17,000 new jobs, not -4,000, as was reported in my previous post. In addition to the underlying data revisions, and correcting for this error, the revised alternative seasonal adjustment for May is -6,000 jobs added (second row of column Wright SA).

Earlier this summer, we examined how the Big Four banks – Bank of America, Citigroup, JPMorgan Chase, and Wells Fargo – performed before, during, and after the 2007-09 financial crisis.  We also blogged about the lending trends within these large banks, expressing concern about the growing gap between deposits taken and loans made by the Big Four, and calling on policymakers to explore the issue further.  We have conducted a similar analysis on the regional banks - The regional banks: The evolution of the financial sector, Part II - and find that these smaller banks are actually faring somewhat better than their bigger counterparts.

Despite the mergers and acquisitions that happened during the crisis, the Big Four banks are a smaller share of banking today than they were in 2007.  The 15 regionals we evaluated, on the other hand, are thriving in the post-crisis environment and have a slightly larger share of total bank assets than they had in 2007.  The Big Four experienced rapid growth in the years leading up to the crisis but much slower growth in the years since.  The regionals, however, have been chugging along: with the exception of a small downward trend during the crisis, they have enjoyed slow but steady growth since 2003.

There is a gap between deposits and loans among the regionals, but it is smaller than the Big Four’s gap.  Tellingly, the regionals’ gap has remained basically constant in size during the recovery, unlike the Big Four’s gap, which is growing.  Bank loans are important to economic growth, and the regional banks are growing their loan portfolios faster than the biggest banks.  That may be a good sign for the future if the regional banks provide more competition for the big banks and a more competitive banking sector overall.