Sources of "The Great Risk Shift"

I want to follow up on Jacob Hacker's recent defense of The Great Risk Shift against misinterpretations of his results relative to the results in a recent CBO report, and more particularly I want to follow-up on the response that his argument has received.

Recall that Jacob finds that household income volatility has been increasing in recent years, but both he and the CBO do not find increasing volatility of individual earnings. It is this difference between the results for individuals and the results for households that many people have focused on (even though Jacob's results are much broader than this one measure of risk as touched upon briefly below).

For example, the response from much from the right is along the lines of this from Rammesh Ponnuru at the National Review Online's The Corner:

Risk Shifts Back, by Ramesh Ponnuru, The Corner: Jacob Hacker has argued that income volatility has been increasing and that Americans are exposed to more economic risk than they used to be, and that liberal policies are required to provide some security. (I bought Hacker's empirical case, but not his policy prescriptions.) The Congressional Budget Office recently released a report suggesting that individual income volatility hasn't been increasing.

Hacker, in response , argues that the volatility of family income has been rising. ... Ross Douthat argues that Hacker has been "taking data that would seem to support the socially conservative contention that changing family structure has had a lot of negative externalities for vulnerable Americans, and using it to claim that 1) Republicans have shredded the safety net and 2) we need a much, much stronger one than what we currently have."

Here's more of what Ross Douthat at The Atlantic Online said:

Risky Business, The Atlantic Online: But the subtitle of Hacker's book, you'll note, isn't "How The Sexual Revolution Created Higher Levels of Risk For American Families and What To Do About It." ...[H]e's taking data that would seem to support the socially conservative contention that changing family structure has had a lot of negative externalities for vulnerable Americans, and using it to claim that 1) Republicans have shredded the safety net and 2) we need a much, much stronger one than what we currently have. The first point is at best debatable... The ... second point just amounts to the rather predictable liberal claim that a bigger welfare state is the answer to increasing family breakdown.

Before addressing these specific claims, let me make a more general point. There has been some doubt expressed about whether the data actually show increasing income volatility. While not conceding that there is a question here to begin with - there isn't much of one and it ought to end soon, see below - here's more evidence. First, there is the Hertz study which makes a strong case in favor of the view that family income volatility rose in the 1990s and thus this work reinforces Jacob's results. Second, there is the Los Angeles Times series “The New Deal”. This series is based on a study that uses different methodology than Jacob uses, but still finds a big increase in volatility since the 1970s just as Jacob finds. [Update: Here's one more study, though I should note it is also preliminary. Its conclusion, based on household-level earnings in the PSID, says "Households appear to have faced notably more uncertainty in the past twenty years than in the preceding few decades."]

Now let's turn to the more specific claims above. The first claim from Douthat is that 1), which says that the safety net has been shredded, is debatable. Jacob's point about public policy seems to be completely lost on most people in this debate. Yet government and private social spending is a quarter of our economy after taxes – it has to be considered as an explanation for changing volatility and risk, and the preliminary evidence suggests that social spending does not reduce family income volatility as well as it did in the past.

The other point that is missed is that Jacob's analysis of household risk goes well beyond income volatility. Thus, looking at income volatility alone, which has been the focus lately, misses a large part of the risk picture and hence misses the extent to which the social safety net has been shredded. Jacob discusses this in the link above.

Now let's turn to point 2) which claims that Jacob's results imply we need to strengthen the family to prevent breakups if we want to reduce economic risk. Many people seem to be jumping to the conclusion that the fact that individual earnings instability isn't the same as family income instability is tantamount to saying that family breakup drives the results.

But, according to the preliminary results, family breakup is not the most important reason for increasing income volatility for households. My understanding is that though the empirical investigation is ongoing, the most likely explanations for increasing income volatility are quite different from the "policy from liberals caused more family breakups, which in turn caused increased income volatility" explanation we are hearing from conservatives. The most likely explanations are:

1. Families rely on two incomes now, so when one worker leaves the workforce, income drops. Likewise, there’s no potential second earner to bump up his/her hours when earnings/hours of the prime worker drop.
2. There are more single individuals. This group has always had higher income volatility.
3. Government taxes and benefits do less to cushion income shocks than they once did.
4. People rely more on asset income, which is more volatile (probably a small effect).
5. There’s more long-term unemployment (also likely a small effect).
6. There’s more self-employment/independent contracting (again, this is likely to be small).

Finally, with regard to this entire debate that has erupted ever since the CBO report appeared, let me repeat and emphasize what I said in an earlier post. Jacob tells me that he is hard at work producing new estimates for public release that take account the many questions he has received on his work, methodological and otherwise, and that the results are holding up nicely and should be ready before too long. If you have any doubt about The Great Risk Shift, I hope you will wait until you’ve seen the new work before coming to any final conclusions.

Comments

[real person from the real world]

Of course employment is a factor. To have a nicely maintained home, comfortably, not to mention children, it takes 2 incomes, now. In IT where the jobs pay better, you have a lot of contracting feast or famine, in addition to imported competition. High paying permanent managerial salaries go to Americans and Green Cards, the rest is between well educated Indians who were vetted thru entry level overseas, who can "hit the ground running" and not so well educated Americans struggling to fit into what's left. THen there are problems with hiring practice: the consultant rejected because he is a "job hopper," someone who was laid off after a merger and took other jobs to make ends meet, now suddenly he is less desireable for taking a sales job after being an infrastructure guy for some tech company and working at big banks, or the position for a techie that can only be filled if you find someone familiar with that proprietary software, or the geek CIO who has his managers mowing thru 40 candidates to find a "hands-on" geek, or the purple squrriel positions that go unfilled and allow companies to say they cannot find Americans, so lets import a few more H1b visas, which more and more benefits a lot of small foreign middleman vendor-employers who act as auctioners for the immigrants that are imported.
Additionally, backruptcy rules have been tightened, there is less and less medical care offered by employers, and for what there is, the prices are artificially ENGINEERED to extract the maximum from consumers, and to benefit various middlemen (insurance).
Can anyone who is trying to maintain a middle class existence say they are not strapped?

[conchis]

I'm a little confused...

If family income volatility is increasing while individual income volatility is not, doesn't that necessarily imply that changes in family composition (essentially family breakup plus family non-formation, your point 2) are primarily going to be driving the difference?

Wouldn't everything else in points 1
and 3-6 contribute at least equally (probably more) to increases in individual earnings variation?

[conchis]

Actually, I guess another way to reconcile the individual and family level trends would be to posit an increased correlation between income shocks within families, either due to changes in matching patterns (e.g. more people marrying others in the same profession) or something else (???).

[ken melvin]

Hardly a surprise. Everywhere we see and hear of less permanence in the job market.

[robertdfeinman]

I still think the psychological aspect needs to be given more study (hard for economists, I know).

If people feel less secure then this affects their behavior, regardless of whether they are actually less secure statistically.

One can easily see this effect in the behavior of those who were of working age during the 1930's. Most of them continued to take a conservative economic outlook for the rest of their working lives and even preached themes of savings and moderation to their children.

It is the third generation which is now in their prime working years that has broken away from this mindset and are the customers for all the new financial products like interest only mortgages and risky derivative type investments.

It would also be interesting to examine the attitudes of those burned in the dot com bust to see if their behavior has changed.

Many young people expect Social Security won't be there for them when they retire. How does this affect their behavior?

[K. Williams]

Like Conchis, I'm confused by this. Your first point -- about having no cushion -- doesn't speak to the difference between family and individual income volatility, because having an earner lose his/her job or stop working would obviously show up in individual income volatility as much as in family volatility. And saying there are more single people, whose individual volatility is greater, may be true, but also doesn't seem like something we should be concerned with, since the point of the CBO study is that individual income volatility hasn't increased. In other words, what you primarily seem to be describing is no great risk shift at all, but rather a shift in the nature of the composition of families.

What am I missing here?

[jalrin]

If one glances at the bankruptcy data that was collected for the Two Income Trap by Elizabeth Warren and Amelia Warren Tygai then you will find that #1 and increasing job insecurity and the rising cost of fixed expenses are the driving forces behind the bulk of this problem which concides with Hacker's hypothesis and runs counter to Douhat's.

What this means is that one job loss can crater the entire household even if the other family member is unaffected (thus answering Conchis' question). Household income can drop for three reasons (brerak-up, wife loses job/gets pay cut, husband loses job/gets pay cut) wherareeas individual income only drops if that person losses their job/gets pay cut.

[reason]

K. Williams,
what you are missing is the falling proportion of income of discretionary. Part of this is voluntary of course - bigger morgages for larger houses. But once some families start bidding up the prices for well located houses, and once certain areas start heading down hill everyone gets pushed into the same trap. Add to this the health care crisis and the rising cost and necessity of college education and the picture starts to emerge.
I'm not sure what is to done about it, but identifying the problem is surely a start.

[reason]

... of income that is discretionary...

[K. Williams]

Reason, I'm not saying anything about how spending has changed. I'm just speaking to Mark's post, and to the reasons for the apparent disconnect between individual income volatility (which has not, according to the CBO study, risen), and family income volatility (which Hacker says has risen).

This applies also to Jalrin's post. The CBO study suggests that "increased job insecurity" isn't, in fact, a major problem, because if it were it would surely have made individual incomes more volatile.

[jalrin]

It has to do with women basically working as much as possible when they are married. When Wives did not work except when they had to, you has a situation where only one individual could cause household variation. Now, look at a siuation where both spouses can be laid off. Individuals may not be more likely to be laid off but (I have not read the CBO study deeply enough to know if they have found something everyone else has missed about all these layoffs), but since two chances for a layoff will affect family income it is more vulnerable then when only the husband's layoff could cause a variation and the wife could mitigate the impact by going to work outside the home (since she is now, it does not boost income).

Read the Two Income Trap. Elizabeth Warren explains this better than I do.

[conchis]

Thanks Jalrin. I think this explanation potentially makes sense, though I'm not entirely convinced by it yet.

I still think that, in order to rely on it, the women you're talking about must actually have experienced an increase in income volatility (in moving from stable lack of income, to volatile positive income*). The question is whether that would have been excluded from the individual volatility data in such a way as to hide the effect. I would assume that would happen if they weren't initially in the workforce at all, but not necessarily if participation increased from a low, but still positive base.

*This potentially highlights the problem with focusing solely on volatility - other things equal, movement into the workforce should be a net gain rather than a bad thing. Though admittedly other things may not be equal.