What If’s in Life
One of the most amazing things about my college class is what one individual, our class President, Jeff Rothbard, has done for our class since our graduation, 50 years ago next spring. Several years ago, after our 40th reunion, he launched a series “Back and Forth” that has enabled many of our classmates to fill out a questionnaire telling what each has done since moving on from school, what they remember most about college, what they are proudest of, and so on. Reading these questionnaires, along with books Jeff regularly asks our class members to recommend, has bound together the lives of our class in a way that never would have been possible had we just gone to the occasional reunion.
But the last question on the “Back and Forth” questionnaire is the kicker and what prompts me to write this post: “if you had a mulligan, what would you do over?”
Most respondents say they wouldn’t have changed anything. I personally sometimes daydream about different things I wish I would have done at different points in my life, but that’s all they are – dreams.
My classmates are an unusual group of people, however. They went to a first-rate school, they are smart, and by and large, they have led successful lives. I am sure all of them have experienced setbacks and unexpected twists and turns that affect us all. But, taken as a whole, they seem pretty satisfied with their lives.
There are at least two reasons for this. One is obvious, and what statisticians refer to as “self-selection” bias. Those who filled out the questionnaires - or “where are we now” statements of life – are most likely to have something positive to say.
But even accounting for this bias there is another reason why so few say they’d like or even have thought about a mulligan. We know best what we’ve seen, what we’ve experienced. It’s hard to imagine anything else, especially to really believe with any degree of certainty how our lives would have been different if we had chosen a different course – a different college, a different initial job or career, a different move to another place, or married someone else.
There are three circumstances in which “what ifs” are not just dreams, but at least in theory, are important to the way we make decisions, in business, in life, and in making policy.
Opportunity Costs
It is textbook economics that when people or firms make investments in time or money – say in going to college or buying a new piece of equipment – they should not only look at the price they pay for the investment, but the opportunity cost of undertaking this investment, and not doing something else. A publicly held company and its management, for example, thinking about investing in a new facility or entering another line of business should consider the financial benefit of spending the money on something else, like buying back the firm’s stock, and only proceed with the proposed new investment if it has the highest expected returns (ideally account for risk). Likewise, high school students thinking about going to college should include both the cost of doing so and the foregone income they could earn by working full-time in the four (or more) years it will take to complete a college degree. For many students, this is a no-brainer: college more than pays for itself over a lifetime. But for others, perhaps specialized training at a community college is a better fit, while sacrificing only two years of foregone income.
That all sounds so rational, so economic. But how many of us weigh all the alternatives before us making any decision? It’s too much work. And besides there is so much uncertainty attached to any choice. Most of us end up doing what Herbert Simon, a Nobel prize-winning economist who taught at Carnegie-Mellon University, won his prize for: “satisficing,” or picking something acceptable, not the optimal or profit-maximizing alternative. https://en.wikipedia.org/wiki/Satisficing.
But that doesn’t mean we should ignore opportunity costs in making life decisions. Implicitly we won’t if we at least weigh at least several alternatives before making snap decisions (I fail at this, though, sometimes, and I’m an economist!).
Counterfactuals
Another term for posing “What if” questions, used often in economics and corners of the law is the “counterfactual”: imagining what would have happened if something about the past had not happened, or when considering a new policy and asking what would have happened instead. Not surprisingly, this also is difficult for many people to do. All we know is the present and what has already happened or may be happening right now. Most of us rarely think about what would have occurred had we followed Yogi Berra’s sage advice, “when get to the fork in the road, take it” – by which, of course, he meant, if we had taken a different fork from one that we did take.
Nonetheless, counterfactuals are used to assess policies all the time. As I describe more fully below, economists for decades have built models of economies. Assuming the statistically estimated relationships that are built into these models hold for some period of time – a big “if” but a reasonable assumption at least in the short run – policymakers, at say, central banks can simulate the impacts of a specific monetary policy adjustment, say raising interest rates, on GDP, inflation and unemployment by comparing them to a “baseline” when interest rates are not increased. The “baseline” is the counterfactual.
Given the increased interest by business in being “climate friendly,” many are attempting to “sell” the reductions in greenhouse gas (GHG) they generate through investments in new environmentally cleaner methods of making things. Firms that want “credit” for being more climate friendly may find the purchase of these credits from other firms less expensive than undertaking their own “green” investments. In principle, these purchases of credits can finance mitigation efforts by those for whom mitigating GHG is the least expensive. The problem, however, is knowing the “counterfactual” – what measures the firms selling and earning money for those credits would have taken anyhow to reduce their carbon measures.
In a recent essay with a graduate school colleague and now professor of environmental sciences at Yale, Rob Mendelsohn, and an entrepreneur colleague with whom I have written some other pieces, John Fleming, I have recently outlined two ways to solve this counterfactual problem. So that the climate credits bought and sold in the marketplace actually reduce greenhouse emissions, rather than constituting marketing ploys to convince consumers and political leaders, among others, that they are making a difference. https://www.brookings.edu/research/a-framework-to-ensure-that-voluntary-carbon-markets-will-truly-help-combat-climate-change/.
I could go on by providing other examples where counterfactual thinking is used in the real world – and not just in dreams – but I hope these two suffice to make the point.
New Policies
Government policy making is or least should be guided by “What if” thinking a bit more than it is. Consider the current debate over the Build Back Better (BBB) plan. Should the federal government make the child tax credit permanent, expand family leave, and so on? How should the government pay for these measures? Advocates claim that the “human” spending initiatives will induce or make it possible for more women to work, increasing their incomes while further boosting the economy. Critics assert that the proposed federal spending, specifically on the child-care credit, will discourage work: why take a job when you can get a government check? A similar critique was lodged against the expanded unemployment insurance payments by the federal government that have recently expired. As for the proposed tax increases, which seem to change by the week, advocates claim that because they will be confined to high income Americans they won’t impact the economy that much, might even improve it. Critics contest that, claiming that many of the rich are like golden geese that lay golden eggs – produce jobs and income – for the rest of us.
Each of these claims are empirical matters. And yet not much attention is paid to any evidence on either side. Part of the reason, of course, is that new policies, by definition, seek a change whose effects cannot be modeled because the change hasn’t bee implemented. But at least when society makes these changes, which often now are made “temporary” to fit under ten-year budget caps, shouldn’t we put in place efforts at least to study the impact of these changes before they expire, and debate resumes on whether to renew or change them? Whatever happens to BBB, shouldn’t some small part of the money be set aside for such studies, however imperfect?
Only over the past couple of decades have economists and other social scientists who engage in this type of empirical research become the leading lights in their academic professions. When I was training to be an economist, and during several decades thereafter, economic theory was all the rage, and those “macroeconomists” who studied and tried to predict wider economic trends – in economic growth, unemployment and inflation, for example – were the kings of the hill in the profession. But even they knew they were limited in what they could do. One can’t run experiments with entire economies and then figure out what to do next.
The next best thing – really the only thing – was to look at historical data and then tease out through increasingly advanced statistical techniques which variables were driving change, and to what extent. Models of the entire economy were built on these statistical relationships. But everyone knew at the time, which has become so evident in recent years, that these historical relationships may only be valid in the short run, as just noted. They would provide little or no guidance about future economic events if those past statistical relationships no longer held. That is essentially what happened with inflation and interest rates over the past couple of decades – both fell and stayed surprisingly low – until now. Inflation is back up, but who knows for how long? While interest rates are still quite low. Economists and financial market experts continue to debate reasons why.
In any event, the dominance of macro in the economics profession has ended, perhaps because the models that macroeconomists had developed have had limited shelf life. In the meantime, a new breed of empirical economists has arisen and taken the profession by storm: instead of building complex, statistically-based models, they analyze all kinds of data to assess the impact of policies adopted in the past, or to infer likely impacts of policies proposed for the future.
One model for how some of these new empirical economists do their forward-looking work is based on how the FDA decides whether to approve a new drug or a vaccine. Although it didn’t do this initially, in recent decades the FDA has insisted on what are called “randomized control trials” (RCTs) that compare first the safety, and then the efficacy, of potential drugs in a group of people who receive the drug (the “treatment” group) with those who don’t (the “control” group). People are randomly assigned to either the treatment or the control. If you’ve ever participated in a clinical trial for a new drug, you’ve seen this first-hand, and probably wondered which group you were in.
By randomizing which group people fall into, all other factors that could conceivably affect the outcome are controlled for, since it is likely that those other factors also randomly affect both groups.
Social scientists, not just economists, have made increasing use of RCTs to measure the impacts of all kinds of social or policy “interventions.” Examples include changes in the way kids are taught, the educational impacts of public “charter schools” (some are positive, others not so); in whether the availability of health insurance makes a difference to people’s health (not as clear as one might think), and recently, whether no-strings-attached cash grants, essentially pilot programs of a “universal basic income” or UBI, affect incentives to work (so far, the evidence seems to be that they don’t). If you want to read more about RCTs in social science, here’s a great non-technical book written by an Australian economist, Andrew Leigh, Randomistas.
There are several problems with RCTs, however. One is their potential high cost. Another is that it can take a long time for the effects of the intervention to show up – Head Start programs being a good example. A third problem is that while pilot studies are taking place, the political momentum for taking bold action may pass.
Given the shortcomings of RCTs, economists have been on the prowl looking for “natural experiments” in society that do not require an elaborate study to be set up and then administered first. The Nobel prize in economics this year was awarded to three leaders in this new branch of empirical economics, David Card of Berkeley, Josh Angrist of MIT, and Guido Imbens of Stanford. I focus on one whom I personally know, David Card, to make the point that natural experiments can produce some surprising outcomes.
Card, with his late co-author Alan Krueger (who chaired President Obama’s Council of Economic Advisers for a time), found in one of the most cited studies of recent years that the difference in the state minimum wage between New Jersey and Pennsylvania allowed them to test whether that difference had an effect on the rate of hiring of minimum wage workers in fast food restaurants close to the state border. Answer: it didn’t, suggesting that at least modest increases in the minimum wage would not discourage hiring. https://inequality.stanford.edu/sites/default/files/media/_media/pdf/Reference%20Media/Card%20and%20Krueger_2000_Policy.pdf.
The other point about natural experiments is that they demonstrate that one doesn’t need a full scale RCT to find interesting things, some of which although having happened already, may inform future policy making. Economists just need to look carefully for natural experiments and see if data are available to test things about them.
The bottom line: we need a lot more of “what if” thinking backed by evidence. It would ground decisions in fact, not slogans, while taking out or reducing some of the uncertainty and emotion about how we all behave, individually and collectively.