Friday, February 12, 2021

Measuring Poverty

    A single parent and worker who makes barely enough to get by and provide for their kids may benefit from receiving health insurance coverage or a housing subsidy. In the case they do receive health insurance coverage because of some governing entity, I can individually make the assessment that they're probably better off. It's reasonable to assume that by covering certain costs, the coverage will increase their general welfare. However, "official" poverty measures, which are based on income and often misreported income do not reflect increases in welfare in many circumstances. In this case, the worker receiving health coverage will not see this reflected in official poverty measures of income; However, these changes would be present in poverty measures of consumption. Income-based poverty measures tend to not adjust for inflation and capture demographic, such as the college educated, who might not fit in the basket of "people we need to help the most." The picture painted by income based measures of poverty are incomplete and adjusting for its shortfalls paints a better picture for analysis and shows the effectiveness of anti-poverty programs.

    Illustrating a complete picture is important as it allows us to create programs that better identify issues and push back against false narratives such as an increasing welfare state hindering the well being of people in the United States. Official poverty is "determined by comparing the pre-tax money income of a family or an unrelated individual to a predetermined poverty threshold." (Meyer, Sullivan 2009) There are many angles we can use to attack how we determine poverty, much more than a single blog post warrants, but one of the main critiques is what pre-tax money income captures. Our single parent worker's health coverage wouldn't be captured by pre-tax money; nor would an EITC expansion or the free lunch offered to help his child. In fact, pre-tax money wouldn't include any noncash benefits which have been the basis of many programs since the Great Society. These noncash benefits should be included in these measures as they reflect changes in consumption; they "have an important effects on the resources available for consumption." 

    One old study that expresses this flawed logic is Hoynes et al. (2006) They found that "despite robust growth in real per capita GDP over the last three decades, the U.S. poverty rate has changed very little." The problem with their conclusion is their use of non-elderly pre-tax money income poverty - for all the reasons stated above. They found that macroeconomic conditions and employment are only responsible for a one percent point decrease in official poverty from 1980 to 2003. In fact, once you incorporate the changing employment of women, poverty rises over this period. Their research "shows" that the falling share of married couples shows a strong decline in poverty and that anti-poverty programs don't aid these people which opens up discussion for them doing the opposite. However, this is far from the case. Once you examine alternative measures of income, once you account for taxes and noncash benefits, once you account for the entire population and not restrict the demographics you're dealing with, once you account for consumption, you get a much different story. It is true, that an increase in single parent families might lead to an increase in poverty, this is both true intuitively and reflected in data, but it is also true that noncash benefits, increases in education and anti-poverty programs do find a decrease in poverty. For instance, investment in education predicts a large reduction in poverty when looking at consumption data, but little to no effect when looking at income data. Indeed, Bruce Meyer and James Sullivan find that when "[you move] from traditional income-based measure of poverty to a consumption-based measure, and, crucially, [adjust] for bias in price indices [you find] that the poverty rate declined by 26.4 percentage points between 1960 and 2010, with 8.5 percentage points of that decline occurring since 1980." A large portion of this decline can be attributed to social security and changes in tax policy such as the EITC. Other programs tend to be less beneficial, but you can also attribute those issues to misreporting and overall design of said programs that create narrow paths of families and individuals receiving said benefits. (Just give people money!) 

    Going off of what I've written above, there are usually many angles you can "uncover" to show the benefits of cash and noncash benefits. There has been much ado lately about a child allowance, a policy I think would do a lot of good in the U.S., not just by making people better off, but by easing political tensions by addressing economic distress. For instance, a child allow is an extremely effective way to reduce child poverty with estimates of a 200-300 per mouth allowance reducing child poverty by nearly 50 percent. There is also the political appeal that people across the political spectrum can endorse. The cost of raising a child is one of the main reasons why people have abortions; by expensing the cost of raising a child we can reduce the number of abortions that people have. A robust approach to welfare can do a lot of good and we should be confident in the socio-economic benefits.

    Cuts to programs such as the AFDC during the 90s "welfare reform" created unintended consequences. This study finds that, although "welfare reform may have had favorable effects on [the] social behaviours of mothers( at least in terms of reduced crime and increased civic participation in the form of voting), the intergenerational effects on social behavior were not favorable, particular for boys, and may have hindered the affected youths' socioeconomic trajectories." These negative social behaviours included "decreased volunteering among girls, increases in skipping school, damaging property, fighting among boys, and increases in smoking and drug use." Many of these reforms were made on the basis of being "pro-family" yet outcomes of those policies seem to reject the goals. The cases where we see the most benefits across all demographics, including strengthening families, is when we have a more robust approach to welfare policy. 

    If I don't think of a segue, I just don't end up doing one. It's my blog! So, at the end here, I'm just going to address a fallacy about the welfare state. There's an argument that the welfare state "encourages families to split up" that I don't find very convincing. It's been dying out, but the general rhetoric still pops up from time to time and I think we should formally eliminate it form our list of primary concerns. There used to be a strong correlation between the robustness of welfare and single motherhood. Something something endogeneity...

Let's take two assertions. 

- Regions with a relatively high number of single mothers are more inclined to pass generous benefits that help those people.

- Regions that culturally accept single mothers are more inclined to pass generous benefits that help those people.

These might (as I say sarcastically; it definitely does!) cause a false positive correlation and if you adjust for fixed effects, this relationship disappears for white people, but remains for black people. Hoynes then found that the relationship between the robustness of welfare and single motherhood in black families disappears when you account for migration; Robust welfare doesn't cause families to break up, but people more likely to be or become single mothers will move to areas with robust welfare policies.

How rational of them...

Saturday, January 23, 2021

Coffee Beans, Copper and Monetary Policy, Oh My!

NGDP targeting? Inflation? Nominal Wages?
Let's take a step out of the academic world. In my opinion these policies are functionally similar and the competence of the monetary authority is what is at play. Establishing a make-up policy is more important.
Say the monetary authority determines the interest rate,
Interest Rate = 4 + 3(inflation - target) + 1(output gap)
Let's take one step back into the academic world. Under an inflation targeting regime, the output gap coefficient would be zero. Now let's take one step out of the academic world. Our monetary authority wouldn't completely ignore the output gap. They wouldn't ignore any piece of data that they think would relate to their decisions to execute a "sound" monetary policy. Hence the phrase, "flexible [insert target] target" where the monetary authority will look to a wide variety of variables to determine the current stance on monetary policy. 
Perhaps you're the monetary authority in Costa Rica, the price of coffee or bananas may be of importance. Perhaps you're the monetary authority in Chile and you want to put weight on the price of copper. Perhaps you're the monetary authority in Singapore and you want to put weight on the exchange rate.
This is often a critique against restrictive monetary regimes, as done by Svensson here on NGDP targeting. 
Recall the equation above, 
Interest Rate = 4 + 3(inflation - target) + 1(output gap)
A nominal GDP target would replace the output gap term with an output growth term and the coefficients on inflation and output growth would be equal - this puts weight on the variable measuring real activity in contrast to an inflation target. Let's take a step back into the academic world. In his paper, Svensson contends that NGDP is too restrictive and while concerns may be valid these concerns aren't functionally different to me once we take a step back out of the academic world. While we're still in, the NGDP target restricts you from looking at variables that aren't inflation and output growth and says that you put equal weight on both of them. As said above, you might want to be flexible, you might want to care about other things and a flexible inflation target would be better. Now, with a foot back out of the academic world, all central banks are always "flexible." Central banks will always take in a large amount of data and go through a process of determining the stance of monetary policy. A restrictive rules based approach isn't going to happen. Even given this, it isn't clear to me that the differences in targets would actually produce drastically different results when observing monetary choices and the competency of the monetary authority at a given point in time.
NGDP still has value:
- Monetary policy is, in large part, about expectations. The relative simplicity of NGDP, or better phrased, "targeting a growth in nominal income", makes setting expectations easier. I mean, just listen to Fed officials talk about the future under it's current inflation targeting regime.... Set up the actual NGDP path and the target path and communicate what you're doing with rates relative to the gap. The two values under an NGDP target are easily observed which contrasts with the unobservable values put into the Fed's reaction function. 
- Selgin and Sheedy write about market efficiencies that would be associated with a NGDP target. Selgin writes about "credit-debtor justice", and in his paper he summarizes how a NGDP target would increase allocative efficiency by allowing for a more counter-cyclical policy while an inflation target would decrease it. Sheedy shows that various households do not have access to financial tools that would help them deal with unexpected shocks to their nominal income. (Something, something 400 dollars...) Firms and various industries do have access to these financial tools and they can equity finance instead of debt finance. A monetary policy that allows for a more counter-cyclical inflation, such as a NGDP target, aids this issue by decreasing the real interest rate on household debt when real output is low.
- By giving inflation and output equal weight, NGDP targeting signals what to do during a supply shock. Other monetary regimes make this less obvious, The simplicity and clarity of equal weights has a bigger priority, in my opinion, over small changes to find theoretically optimal tradeoffs, especially since NGDP is probably "good enough", in many models. Also, models play off the real world, and since we don't perfectly model the real world... we're fine! Factor this into the fact that central banks are inherently "flexible" as said above, and you see why these policies are functionally the same with one step into the real world.
Now, any regime has critiques and I think the critique below of NGDP targeting is valuable, but fixable. (Politics aside)
NGDP Growth = Inflation + Real Output Growth
We know that real output growth tracks productivity growth over the long run. (Btw, listen to this!) If productivity growth is low over a decent stretch of time, for any given NGDP target, inflation would be high. However, we can also change the target every few years. We might set a rule where we update our target every X amount of years based on productivity and overall macro trends.
Recall our equation above, 
Interest Rate = 4 + 3(inflation - target) + 1(output gap)
A level target would replace all growth rate terms with "level-deviation" terms and it's easy to see why a theoretical make-up policy has bigger implications in the real world.
The Fed's new AIT isn't exactly a level target, but it is a form of make-up policy and we're the implications of that right now.
Now, what's the real point of this post? Everything and nothing. Lots of these distinctions don't matter. Some of these distinctions can matter more than others. Really, the point of this post is just me saying this:
- The target in of itself probably isn't functionally different in a real world despite many academic debates surrounding it. While different targets allow for different things to be "seen" better or worse, these things can be done under any regime given that central banks take in a large amount of data.
- While these distinctions may not be different by much, NGDP targeting has a lot going for it over other options and some distinctions may be more relevant than others. Despite this, some of the issues done under inflation target didn't have to happen even if NGDP makes them "less likely" by being a better lens to see the stance of monetary policy through.
- What really matters is a form of make-up policy to address monetary concerns by making up for past mistakes.
- Also, I just wanted to post something on my blog. :)