Following the Trump administration’s decision to separate families, I’ve felt a mix of sadness for the families, anger at those directing & facilitating this policy, and troubled by the stories of neglect and abuse coming out of the childrens’ “shelters.” Like many, I was moved to donate towards family reunification efforts but eventually opted not to. My intention in writing this is not to discourage anyone from donating towards important causes, but to illustrate that some problems cannot be solved by fundraising more donor dollars.

Without getting into the history, details, or politics of immigration policy, a major barrier preventing separated families from reuniting was the detained parents’ lack of money to post bail. If the arrested parents could pay for a bail bond, they could be released and reunited with their children. Immigration bail bonds reportedly cost between $1,500 and 10,000 and several organizations fund bail for detained immigrants.

The Solutions
However, after reading up on immigration bail bonds and reading about the viral fundraising successes, it was clear that more than enough money had been raised by dedicated immigration bail bond funds. If there were only one kid per family (roughly 2,500 kids) and the maximum allowable bail of $10,000 was set in every case, it’d only take $25M to bail out at least one parent from each family. Some organizations raised that amount single-handedly. In short, additional dollars donated would have no impact on the issue of family separations (although the early donors may have had a marginal impact and all donations can hopefully be used to fund bail bonds unrelated to family separations).

The second area I looked into was legal services. It’s difficult for residents of the US to understand the legal system and even harder for a new immigrant (and especially under the duress of being separated from their children)! However, my back-of-the-napkin analysis also indicated that the organizations doing this work had raised more than enough to hire sufficient legal personnel. The primary problem does not seem to be lack of funds or capacity for hiring enough people, but the government has not properly documented all the people that they’ve detained and likely deported many parents without their child(ren). It has and will continue to take time to reunite families and it’s difficult to see how more money will expedite the process at this point.

What I Did Do
While I declined to donate, I did advocate in small ways; posting on social media, mentioning here, joining a local rally with my family and friends, and continuing to advocate and donate to organizations that attack the root problems in Central America. Trump created and enacted the policy and had the power to reverse it, which he eventually did when it’s existence and details became known and deeply unpopular even within his family, party, and among his staunchest supporters. In this case, media attention, public scrutiny and advocacy, and legal action were the influential factors.

Should I donate at all?
The point of this post is not “don’t donate.” I believe in giving and giving generously, but giving money isn’t always helpful (and can sometimes be harmful). I find that my initial reaction to donate is often based on a desire for some sense of agency (“don’t just sit there, do something!”) or driven by my ego (“good job Matt, you’re helping out”), which are not good reasons to donate. I believe that if we ignore false narratives and selfish desires, then we can be more efficient and effective with what we do give. At some point, I should probably post some case studies of instances where I did donate, rather than solely focusing on times that I declined. I believe that we should give, but thoughtfully.

Although the headline numbers on last week’s job report were below expectations and tariffs may create further drag, the labor market remains quite strong. One of the metrics that I have been watching has been the continued decline in unemployment rates among groups with historically higher unemployment. While all jobs and compensation are not equal, low unemployment rates are an encouraging sign.

One way to look at the breadth of the labor market’s health is to look at unemployment rates by education level (see below). As the bottom panel shows, the difference in unemployment rates between those with a bachelor degree and those with a high school diploma is historically low.

Similarly, we can slice the data to view unemployment rates by race. Currently, the gap between black and white unemployment rates is also historically low.

Another metric worth following is the U6 unemployment rate, which includes the official headline U3 unemployment rate plus those who are underemployed (discouraged workers or those marginally attached to the labor force, as well those working part-time for economic reasons).

While low unemployment benefits workers today, it also provides valuable training and experience that will benefit those workers, their future employers, and the overall economy for years to come and beyond the current cycle.

Negative screening (or exclusionary screening) is the process of screening specific assets out of an investment universe or strategy. Implementing these screens is called divesting and results in divestment.

Divesting does connotate excluding something that one would otherwise own. For instance, a large-cap fund might not own tobacco stocks or small-cap stocks. Intentionally screening out tobacco stocks would be considered divestment, because it is something the fund would own otherwise. However, the fund would not be considered to have divested from small-cap stocks though, as they fall outside a large-cap fund’s investment universe. In other words, divestment is the intentional act of excluding assets that one would otherwise own.

There are purely economic reasons to divest from assets, but I will focus on values-based divestment here. Values-based reasons to exclude assets may include not wanting to be associated with or derive any benefit from specific activities/products/services. Values-based screening has ancient roots as various religions have forbidden debt with very high interest (or any interest at all) for thousands of years (see here). More recent examples include the Quakers divesting from the Atlantic slave trade or US investors divesting from Apartheid-era South African assets. Today, hot topics include tobacco, weapons, and fossil fuels.

It should be noted that there are arguments that even non-economic considerations can also be economic ones. For instance, as energy consumption shifts towards renewable energy, the returns from fossil fuels may suffer. A mass shooting may bring unwanted publicity and legislation upon a firearm manufacturer. These are cases where values issues may become value issues.

There are a wide variety of objectives and approaches to divestment, each with a unique set of benefits and drawbacks. Exploring all of the applications of divestment is beyond the scope of a blog post, but we will look at some common examples in the coming weeks.

An important correction/update to my last post.

Upon further analysis, I found that the low-emission portfolio’s overweight to companies with bottom-decile ESG scores was the result of including companies with no ESG scores. The above chart has been updated to only include holdings with an ESG rating. As you can see, the result is that the low-emission portfolio scores very similarly to its benchmark. This illustrates that investors need not sacrifice ESG qualities to materially reduce emissions exposure. On the other hand, the portfolio still has benchmark-like ESG scores, rather than broad improvement. Thus, I believe the main point that investors must prioritize what exposures they want (it may be impossible to target a portfolio with high ESG scores, low emissions, gender diversity, and low tracking error) still remains the same.

6/27/2018: IMPORTANT UPDATE AND CORRECTION TO THE BELOW HAS BEEN POSTED HERE: http://thoughtfulfinance.com/2018/06/27/update-trade-offs-high-level-esg-scores-vs-specific-tilts/

Now that we have looked at how ESG considerations and index construction methodologies can impact the ESG characteristics of indices, it is time to delve deeper into specific issues. For instance, rather than optimizing high level ESG scores, an investor may want to primarily optimize the environmental (E) and governance (G) scores without regard for the social (S) score. Or an investor may want to focus on more specific or objective issues such as lower emissions or gender equality. Again, we will be looking at a single fund that is well-known, tracks a well-known benchmark, and is sufficiently scored by a well-known ESG analytics firm. Not comprehensive a study, but illustrative of some basic dynamics.

This particular fund targets low greenhouse gas emissions and avoids owning fossil fuel reserves, while also divesting from stocks with severe controversies. The emissions metrics look great as shown below:

	Fund	Benchmark
Tonnes of CO₂ per $1M invested	37	103
Tonnes of CO₂ per $1M of sales	58	206

Source: Bloomberg, FossilFreeFunds.org

However, the fund’s distribution of ESG scores tells a different story, as shown below:

The fund (green area) is overweight companies in the lowest-scoring decile of ESG scores and underweights nearly all of the better-scoring deciles (relative to its benchmark index, the red area). Thus, this particular fund appears to achieve its emissions goals at the expense of its high level ESG score. What is an investor to do?

Some investors may want to take a balanced approach and modestly improve the high level ESG scores of their portfolio. Other investors may want to focus more on emissions or gender equality or some other issue, using a fund like the one above. It is also possible to act in a more nuanced way by blending the two approaches, although this can be difficult with the existing universe of ETFs and mutual funds. A word of warning to do-it-yourselfers though: consider the above tradeoffs when constructing a portfolio, as bolting together different funds could result in unexpected exposures (ie. a low carbon fund might more than offset the positive ESG characteristics of an ESG-optimized fund).

There are no perfect portfolios. Just as traditional investors must prioritize returns, risk, volatility, liquidity, and so on, ESG investors must prioritize what factors are important to them. There is no “right” way to invest, nor is there a “right” way to invest responsibly. Investors must determine what is important to them and then decide which trade-offs they are willing to make.

The below comparison is obviously not a comprehensive study or representative of the many inherent nuances, but I do believe it illustrates something that I frequently find when evaluating off-the-shelf ESG investment products (such as mutual funds and ETFs). Below are two large-cap index ETFs from the same sponsor using the same ESG rating provider.

• The first fund (orange) screens out a minimal amount of industries and then weights companies based on their respective ESG ratings.
• The second fund (grey) screens out a much larger number of industries and then simply weights the remaining funds by market cap.

The distributions look pretty similar. The first fund (orange) primarily shifts allocations from decile 10 towards decile 8. Is this materially different or better than the second fund (grey)? I would argue no. The mode, median, and mean may look different, but the underlying scores are not necessarily better or worse. Here’s the same data graphed in a different format:

I’ve looked at many SRI/ESG index products that use various screening and weighting methodologies, but I have not found major differences in the distribution of ESG scores between funds that take an exclusionary approach versus the ones that take an optimizing approach. The net effect of either approach is basically to shift the quality of ESG factors to the right. There may be minor differences around the edges, but I have found the distributions to be largely similar. Again, this relates primarily to index-based strategies.

What does this mean? It means that there are multiple approaches to improving the ESG characteristics of indices. Excluding the “bad” stuff or simply underweighting it often produces similar results. Both approaches can shift a distribution of ESG rankings similarly. Investor motivations, constraints, and objectives will determine the best approach to take (and I don’t advocate one over the other). Ideally, investors can integrate both approaches, although this can be a bit more difficult to execute well.

Two of the main arguments made by ESG advocates are that  “values can impact value” and “ESG factors correlate with better corporate financial performance.” I’m sympathetic to both arguments, although they both seem better suited to an active management context (due to higher levels of portfolio concentration and greater emphasis on fundamental analysis). Yet, as investors increasingly shift assets from active to passive, it is worthwhile to examine how ESG considerations impact index characteristics.

The below is not meant to be a comprehensive study, but simply illustrative of what investors can expect from a plain-vanilla ESG-oriented portfolio. The red area (below) represents the distribution of high-level ESG scores (from a well-known ESG analytics firm) for the portfolio of traditional large-cap index fund. The green area represents the same distribution, but applied to an ESG-themed large-cap index fund that optimizes for ESG score.

This particular fund is fairly representative of many ESG-themed funds and receives good high level ESG scores from multiple ratings firms. While the distribution is shifted meaningfully to the right, it is still a relatively modest shift. This is not meant perjoratively as radical shifts often result in unacceptable risks and/or tradeoffs.

Below is the same data in a different format, but with the same conclusions: there is a slight shift of weighting from companies in deciles 2-5 towards companies in deciles 8-10.

When investors are optimizing for high level ESG scores, it is important for passive ESG investors to understand the above dynamics and to maintain reasonable expectations in terms of objectives, portfolio characteristics, and performance. This is especially true when considering the uncertainty surrounding high level ESG ratings (see here and here), the increasing amounts of “greenwashing,” and questionable claims from some sponsors and managers.

Recently came across the below chart, which illustrates the weak correlation in company-specific ESG ratings from different providers (which I wrote about earlier this week). Even though the below uses slightly different data sets than my previous post, the R² is remarkably similar. I’ve heard other experts reference correlations in .3 to .4 range, which seems consistent with the below and my previous post. I believe the takeaway is the same: the scores are directionally similar (the below dots generally move from bottom left to upper right), but the correlation is relatively loose.

Two weeks ago, we looked at Facebook which is included in many ESG-themed indices and funds despite corporate governance concerns and a recent data privacy scandal. Some ESG managers even divested from FB following the scandal. Although Facebook is the most recent example, there are divergent views and competing opinions regarding the ESG characteristics of many companies.

A quick glance at some of the largest ESG funds reveals that there is a range of opinions regarding many companies, from JP Morgan to Exxon Mobil to Netflix. Some ESG funds own these names and some do not. We can see this visually by pulling ESG ratings from two different providers for companies in the S&P 500 index. The results (below) are all over the place with a weak correlation. Upon first glance, it appears that ESG factors (like other investment factors, such as quality or value) are in the eye of the beholder.

However, things look much different at the fund level. Below are the respective ESG scores from two large ESG ratings firms for some of the largest large-cap ESG index ETFs. The correlation is much higher and the scatterplot falls within a much tighter range (blue lines). Sure, there are differences in methodology, holdings, and weights, but the aggregate ESG ratings are similar.

This is similar to relationships that we find in the traditional investment world. For instance, value funds weighted by P/E, P/B, P/CF, dividends, CAPE, and so on, may all have different holdings and weights, but they largely share similar characteristics and factor tilts.

Just as investors should not get too hung up on the high level numerical ESG scores or use subjective judgments arbitrarily, it does help to understand the inputs, assumptions, and overall methodology of the ESG ratings. The ESG ratings of individual companies can vary wildly, but a larger portfolio (such as a mutual fund or an exchange-traded fund) will diversify a lot of that variability away (as shown above). However, funds are constructed in a variety of ways and there are both better and worse ESG funds out there, so investors still need to look under the hood to know what they own.