Deepseek gets F+ in Econ

LLMs have trouble with this pretty-basic economics question:

Prompt: “Is this factually accurate?

Positively correlated price-quantity pairs reveal that the demand curve shifted.”

 

The LLMs will say that this can be false, but always without offering a counterexample or even mentioning that a counterexample could be the basis for such an answer.  OpenAI’s o1 model is sometimes an exception, stumbling toward a realization that no counterexample exists under the usual assumption about supply and demand slopes.

 

All of OpenAI’s models provide an excellent answer if only given a little hint to use logic.

 

Prompt: “Is this factually accurate?

Positively correlated price-quantity pairs reveal that the demand curve shifted.

[Hint: consider proof by contradiction]”

 

Grok2’s and Deepseek’s answers do not improve with the hint.  Grok2 shows a lot a confusion about what is a proof by contradiction.  Deepseek is the worst kind of student, giving a long-winded incorrect answer laced with confidence:

GPT-4o gives a nice answer when provided the hint:

Analysis of Positively Correlated Price-Quantity Pairs

The statement that "positively correlated price-quantity pairs reveal that the demand curve shifted" needs to be carefully examined. Let's consider the dynamics of demand and supply to understand the accuracy of this statement.

In basic economic terms, the law of demand suggests that, all else being equal, as the price of a good decreases, the quantity demanded increases, and vice versa. This creates an inverse relationship between price and quantity, typically resulting in a downward sloping demand curve.

However, the existence of positively correlated price-quantity pairs (i.e., both price and quantity move in the same direction) indicates that other factors could be at play. This situation can occur if there is a shift in the demand curve itself, rather than a movement along the existing curve.

Proof by Contradiction

We can analyze this using a proof by contradiction:

1. Assume that the demand curve does not shift, and thus only a movement along the existing demand curve occurs.
2. If the demand curve is downward sloping, an increase in quantity demanded should occur only with a decrease in price, reflecting an inverse relationship (as per the law of demand).
3. If price and quantity are positively correlated (both increase or both decrease), this contradicts the inverse relationship expected with a movement along a downward-sloping demand curve.

Therefore, the only logical conclusion is that the demand curve itself must have shifted (due to changes in consumer preferences, income, number of buyers, expectations, etc.), allowing both price and quantity to increase simultaneously.

Conclusion

Thus, the statement is factually accurate: positively correlated price-quantity pairs suggest that the demand curve has shifted, rather than indicating movement along an unchanged curve.

---

---

With that said, GPT-4o requires 6 seconds to handle this easy question.  In contrast, for almost 10 years my automated economic reasoning system has been able to answer such questions with 100% reliability and 100X the speed.

Biden-Harris policies and their consequences were no surprise to those paying attention

Milton Friedman used to advise researchers to focus on large policy changes rather than attempting to separate a small change’s signal from the noise. In this sense, the “ambitious” policy agenda of the Biden-Harris administration was expected to be a gift to the research community.

Accepting this gift, since 2020 I have been making forecasts of some of the consequences of those policies. Now is a good time to assess the accuracy of those forecasts, which relate to aggregate labor markets, insurance price controls, and drug price controls.

As promised, Biden and Harris redistribute income with health insurance expansions and student-loan forgiveness, although not necessarily in the Robin Hood direction. They give union bosses more tools for reducing competition in the labor market. They try to regulate the internet as a public utility. They distort healthcare markets in many ways, including a new ban on short-term health insurance plans, and granting selected companies a monopoly on a generic drug. They go significantly further than the Obama administration in terms of requiring the private sector to change behavior in ways the bureaucrats expect to reduce carbon emissions. At great human-capital expense, they enabled teacher unions and blue-state governments to maintain “social distancing” far longer than warranted.

The exhaustive list would have more than 1,000 entries.  Overall, even the federal agencies’ own low-ball estimates of the costs of the regulations finalized 2021-24 are almost $2 trillion.

1. Macro Performance

Four years ago, I released a study with Kevin Hassett, Tim Fitzgerald, and Cody Kallen of the economic effects of candidate Biden’s agenda compared to President Trump’s. Knowing that campaign promises do not necessarily turn into policy, we analyzed several policy scenarios. The scenario closest to the portfolio of policy changes over the past four years we called “capital taxation constant” (CTC). Biden-Harris climate regulations proved to be somewhat more aggressive than represented by the CTC scenario including, for example, a requirement that manufacturers of medical inhalers (sic) either cease production or convince the Environmental Protection Agency that they are earnestly seeking lower-emission technologies. On the other hand, as nonlawyers we did not account for such a high failure rate of Biden-Harris rules in federal courts.

Under the CTC scenario, labor and capital would be 5.0 percent below the Trump baseline in the long run. In tomorrow's Wall Street Journal, we show that a single trend fits the data well from 2017-Q1 through 2021-Q4, except for the first full pandemic quarter. Then inflation hit and employee compensation—and national income more broadly, which isn’t shown in the chart—fell 5 percent behind. To be more precise, the latest data (2024-Q2) show inflation-adjusted employee compensation per adult to be 4.6 percent below the trend.

Arguably, human capital would have fallen somewhat below its trajectory after 2020 due to pandemic behaviors unrelated to Biden-Harris policies. By itself, this would have pulled labor income somewhat below its prior trend for several years. On the other hand, with some of the regulations not taking effect until 2025 and beyond, we have not yet seen the full effect of the Biden-Harris policies. Large language models and other “AI” technologies have been a positive growth effect that was unanticipated in 2020 when we made the forecasts.

We used a closed-economy model (tariffs were modeled like other excise taxes) that imposes a constant labor’s share, a constant depreciation rate, no statistical discrepancy, and equality between GNP and GDP. In reality, labor’s share of national income has been pretty constant, but the national income’s share of GNP has fallen a bit. More significant has been a fall in the ratio of GNP to GDP. Conversely, a real GDP per capita chart would look “better” than our compensation chart, which of course is no consolation for workers.

Why I have an OpenAI subscription

 It's $20 per month.  Here are the features I value and use:

• Others' custom GPTs
• My own custom GPTs
• Integration into email (Thunderbird) and word processing (MS Word for Mac OS)

In many cases, a custom GPT is nothing more than somebody uploading their documents that become the primary information source.  The creator can also train the custom GPT by giving it queries and then pointing out mistakes or room for improvement.  In some cases, the creator may also work with OpenAI to provide additional capabilities like Wolfram has done.  Regardless, the OpenAI subscription gives access to custom GPTs created by self and, if shared publicly, others.

Wolfram's Custom GPT

The Wolfram GPT, for instance, combines the "verbal skills" of Large Language Models with the analytical capabilities of computational AI.  As needed, Wolfram's GPT automatically calls an algebra engine or other computational engine and then processes the response.  Learn more here.

Since 2016, my coding has been primarily in the Wolfram language. While ChatGPT Plus can be a helpful aid in this coding, my own capabilities surpass it. The Wolfram GPT, however, operates at an entirely different level. This is demonstrated by a time-aggregation function it wrote for me. 

I have used @, /@, @@, @@@, #, #1, #2, :>, ->, but had no idea &-> was a thing!

Here's another example from Wolfram:

My own custom GPTs

I have several of my own custom GPTs, which are created merely by uploading my own documents.  Those docs become the LLM's primary knowledge source.  One of them is "Chicago Price Theory Tutor" which is amazing (but not yet shared with the public :)).  How smart is this?

Custom GPTs like this can understand and repeat back algebra (see the previous Wolfram screenshot).  I did not have to train it in this regard -- just upload a document that had algebra in it.  It is terrible at charts or graphs, so I instruct my custom GPTs not to even attempt to make a chart.

Once the Chicago Price Theory Tutor gave a wrong answer.  But then I realized that the document I provided was correct but confusing on that topic.  The 2nd edition of Chicago Price Theory will explain better.

Integration into email and word processing

The subscription allows you to interact with the LLM via API.  You could write your own code to call the LLM, but all I have done so far is use others' code that I prime with my OpenAI API key.  The Aify add-on for Thunderbird email adds a button to email-compose windows to perform various LLM tasks on the text therein, such as summarize it, check grammar, recommend edits, etc.  "GPT for Excel Word" is an add-on that performs similar tasks in MS Word, and presumably in Excel as well, though I haven't yet tried Excel.

There is an extra charge (beyond the $20/month) for API usage.  The extra charge is proportional to the amount of text sent back and forth to Open AI.  A busy day for me in this regard generates charges of less than $1.  "GPT for Excel Word" levies its own charge of the same order of magnitude.

Accessing the API this way, or other ways, also provides quite a large "context window" that improves results for certain tasks.  I think the custom GPTs also, in effect, have a large context window through uploading sizable documents.

Supply and Demand Curves with the "Wrong" Slope

"Personal Increasing Returns" is a way to describe goods for which consumers can purchase access to a lower price.  An important instance is human capital, where consumers reduce the leisure-opportunity cost of consumption by investing in skills.  Another is consumer financial management, which Sala-i-Martin and I studied in this JPE paper about purchasing access to cheaper future consumption.  The demand for illegal drugs can also be understood in this way.  

It's one model with various variable labels.  A marginal cost curve sloping the "wrong" way unifies the applications.  My purpose here is to compare and contrast this and other non-trivial instances of wrong-sloping supply and demand curves.

The “supply” curves shown in Figures 2 and 5 of "Personal Increasing Returns" are closely related to what price theory has traditionally described as “forward-falling supply curves.”  Examples are Marshall [Book IV, Chapter XIII] and Adams and Wheeler [1952], although they focus on diminishing marginal cost at the industry level rather than the personal level.^[1]  Either way, sufficiently strong complementarities are required. 

[click image to see side-by-side]

Although "Personal Increasing Returns" assumes constant input prices, including the relative price of human capital h, upward-sloping input supply could introduce upward-sloping elements in Figures 2 and 4.  This occurs at quantities where the complementarity between h and q (e.g., human capital and labor supply) is insufficient to offset increasing input prices.  Still, such supply curves are known as forward falling as long as they have a declining portion.  Each quantity has a unique marginal cost, even if each marginal cost does not correspond to a unique quantity.

While Marshall’s formulation with firm-level supply that is a function of industry output may produce a forward-falling industry supply curve, the complementarity could be on the demand side instead.  Becker [1991] posits that a consumer’s demand curve may be shifted out by market-level demand, particularly in markets for “social activities in which people consumer a product or service together and partly in public.”  As a result, the market level demand curve slopes up in the price ranges where the complementarity is sufficiently strong.  In that range, an increase dQ in the market quantity shifts the sum of individual demands even more than dQ due to the social interaction, so price must increase to choke off that additional demand (that is, limit the increase in the sum of demands to dQ).  Still, each quantity on the market-level curve corresponds to a unique price, even though a single price may not correspond to a unique quantity.

The snob-good model is the opposite of Becker’s social interactions in that some of the consumers value the good more when others are not consuming it.  The resulting demand curve may have a backward S shape, with only the elite consuming at the highest prices.  As the price falls enough to induce the general population to consume, enough elites may leave the market to reduce total demand, thereby generating the middle part of the S.  Once enough elite are gone from the market, further price reductions increase market demand.^[2]  Unlike Becker’s demand curve, the snob demand model can have multiple market prices corresponding to a single quantity.

The economics of “backward bending” supply curves and the demand for Giffen goods is entirely different.  In those cases, the usual substitution effects on supply or demand are more than offset by income effects.  For the backward-bending supply curve, the supplier’s alternative to production is consumption of a normal good; the classic example is the supply of labor.  A Giffen good has income and substitution effects in opposite directions because it is an inferior good.

^[1] Stigler [1949, p. 165] is reluctant to discuss much about forward-falling supply without sufficient demonstration of its importance in practice.  Respecting that principle, "Personal Increasing Returns" points to education decisions, vehicle choice, drugs and alcohol pricing, and other situations where consumer action can significantly affect a price they pay.  Boulding [1955] refers to “falling supply price.”

^[2] See also the Veblen [1899] model, algebraic versions of which can be found in Leibenstein [1950]. 

Adams, Robert W. and Wheeler, John T. (1952). "External Economies and the Falling Supply Curve." The Review of Economic Studies 20(1): 24-39.

Becker, Gary S. 1991. "A Note on Restaurant Pricing and Other Examples of Social Influences on Price." Journal of Political Economy 99(5): 1109-16.

Boulding, Kenneth E. 1955. Economic Analysis. New York: Harper & Row.

Leibenstein, H. 1950. "Bandwagon, Snob, and Veblen Effects in the Theory of Consumers' Demand." The Quarterly Journal of Economics 64(2): 183–207.

Marshall, Alfred. 1890. Principles of Economics. London: Macmillan

Stigler, George J. 1949. The Theory of Price. Macmillan.

Veblen, Thorstein. 1899. The Theory of the Leisure Class. New York, NY: Macmillan.

Dublin Chaos Last Night: Firsthand Account

Between 1 and 2pm local time, young school children were stabbed in Parnell square by a single assailant.  We were less than a mile away at John Mulligan's pub at the time; learned about the stabbing on the news during the afternoon.

Not yet aware of events, we went by Parnell Square about 90 minutes later.  A few dozen men aged 20-40 were cursing at the police (gardai).  Presumably they were locals given that several were yelling in Gaelic.

About 3 blocks south of Parnell Square is the OConnell bridge, the one nearest Trinity College, where at 8pm two buses and a police car were set on fire.  We were on Fleet Street, which is about 2 blocks west.  We heard helicopters, which suggested trouble although many tourists and locals were still conducting business as usual.  As we walked down Fleet Street, in the direction of the bridge, a couple of white teen males pushed passed us wearing black face masks and wearing at least 2 empty backpacks each.

I thought about telling the teens that face masks are for criminals.  Undoubtedly they were criminals about to commence looting, which was occurring about that time at about five merchants near the bridge.  The criminals did not look like poor people and certainly were not emerging from a poor neighborhood.

The criminals likely numbered several dozen, which is a small percentage of the number of people present.  The bridge is in the heart of the tourist area and near Trinity College (itself a tourist attraction).  Of course, there are many pubs, some catering to tourists and others to locals.  A fraction of those who had drunk by 8pm were interested spectators.

We saw merchants calmly removing merchandise from their shelves, as if they were instructed to do so or were following a precautionary drill.  Trinity College was being evacuated, said a Trinity College staffer as we briskly walked past.  Students were flowing out of the campus.  I asked a few where they were evacuating to and they said “home, away from here.”  Several minutes later we would learn that our hotel lobby, about 3/4 mile west, was locked as a precaution.

A block west of Trinity College we saw many police.  They seemed organized, but quite nervous.

Note that the fires and looting are unprecedented in recent Dublin history yet the merchants and criminals seem prepared.  This fact pattern suggests to me that maybe law enforcement had intelligence about organized retail crime and in prior weeks had warned merchants about that capability.

The chaos has been attributed to “misinformation” (about the knife assailant's nationality) and “right-wing” political sentiments.  Regardless of any advance warning, this looks like organized crime is a big factor in addition to, or rather than, something political.

We saw a bus burning and smelled smoke throughout the area within a quarter mile of Trinity College.  This was a bit confusing to at least one of our kids, who was worried about “bombing,” because we had just returned from Northern Ireland where we learned about sectarian violence that had ended 20 years ago (Dublin is not part of Northern Ireland, which is thriving in terms of tourism, construction, and more).

The next morning I heard an elected official say that she was eager for the criminals to serve time in jail — not a response I would expect in Chicago.  (Although Chicago as part of USA does provide greater free speech rights—see below).  Reportedly the attacker was detained by unarmed civilians (update: one was a food delivery man who used his motorcycle helmet as a weapon).  Police are typically unarmed, too.

PS.  In Ireland and Northern Ireland it is illegal for “any person in a public place to use or engage in any threatening, abusive or insulting words or behaviour with the intent to provoke a breach of the peace or being reckless as to whether a breach of the peace may be occasioned.” Perhaps that explains why the pro-Palestine protest we saw in Derry was rated G?

"Justicia Social" Failed Venezuela's Indigenous and Mestizo Populations

“Capitalism is the road to hell, to the destruction of the world.  …it’s the freedom to oppress, to invade, to kill, to annihilate, and to exploit.”  Hugo Chavez, 2009.

“Help me become President and I’ll help you get the land back.”  Hugo Chavez, 1998.

 

The quotes summarize lessons increasingly being taught in U.S. grade schools, high schools, and colleges.  Particularly referenced in “social justice” lessons are indigenous peoples, whose representatives also backed Chavez.  The lessons should be considered with robust reasoning, and compared to data.

My purpose here is to consider the case of Venezuela, where Hugo Chavez was President beginning in 1999, as an alternative to “capitalism” for indigenous peoples and their descendants.  At the time of the above quotes, if not now, two-thirds of the population of Venezuela was mestizo, meaning that each had a mix of indigenous and European ancestry.  A related comparison can be found elsewhere: Cuba and Puerto Rico from 1950 to present.

While many teachers of U.S. students dream about trying something other than capitalism, it became a reality for Venezuelans.  Land was redistributed in the name of equity.  So were petroleum assets, also as promised.

 

Chavez shirts sold well on college campuses.  U.S. intellectuals on the left were “mesmerized” by Chavez.

 

One direct result of land reform was violence.  Food became scarce.  No longer satisfied with the assets of “rich people,” the government came for the small businesses too.  Over 7 million Venezuelans have left the country in the past ten years, when the population was about 30 million.

 

This alternative to capitalism, steeped in “justicia social,” did not go well for people with indigenous ancestry.  My comparison here is not with nirvana, but with the capitalism they had before and the capitalism they will get wherever the migrants end up.

 

“Anti-capitalist” and “de-colonial” teachers today rarely mention Venezuela, Cuba, or other real-world alternatives to capitalism.  With only a few exceptions, there is no attempt examine the mistakes of the past so that next time can be different.  Does the omission reveal laziness, or some interest in repeating such experiences?

 

[Image source: https://twitter.com/venanalysis/status/1447981828254617604]

No More Troubles

 

• The violence seemed like it would never end; bombings, shootings, and kidnappings became so commonplace that they were almost expected, a grim part of the daily news cycle.
• Even simple tasks, like going to the shop or sending the kids to school, were fraught with uncertainty.
• The politicians seemed unable or unwilling to bring about meaningful change, trapped in their own ideologies and the weight of history, if not poorly incentivized.

 

This was Northern Ireland during the three decades of the Troubles, itself preceded by 300 years of conflict between Ireland and Britain.  During the Troubles, about 1/500 (cumulative) residents of Northern Ireland were killed by shootings, bombings, civil unrest, and the like.

 

I have been reading Leon Uris’ Ireland: A Terrible Beauty published in 1978.  Although Uris is known for his fiction, this is a non-fiction travel book.  It has lots of bitterness and anger and would not be easy for a British reader to stomach.  It is difficult to understand from today’s perspective.

 

A peace agreement was reached that has lasted 25 years now, even through Brexit.  Today “No more troubles” is a slight exaggeration, but there has been a reduction of almost two orders of magnitude according to the Global Terrorism Database.

Something worked well in Northern Ireland.  U.S. President Clinton and U.S. Senator Mitchell get a lot of credit for brokering the 1998 Good Friday agreement.  Interestingly, nobody seems to remember exactly which part of Ireland Mitchell’s father was from; perhaps that ambiguity helped the negotiations.

Note: per GTD coding, Israel (country code 97) does not include the West Bank or Gaza.

Image Credit: Generated with AI (Bing Image Creator) ∙ October 8, 2023 at 6:48 AM

Private Property: That Spot

Spot is the dog featured in an eponymous short story by Jack London.  Spot’s owners desperately want to be rid of the dog.  They tried shooting him.  There is an attempt to kill Spot with an axe.  They tie up Spot and leave town.  Spot is set lose on a river ice floe.  Spot always finds his way back.

On several occasions, governments (and others) have tried to eliminate private property.  Yet, it finds its way back.  After the Russian Revolution, most of the urban housing stock was nationalized.  The government used its authority over those properties to punish political opponents as well as those who did not perform well in their state-appointed jobs.  But, as documented in this UChicago History dissertation, still tenants traded apartments, improved them, and passed them on to heirs.

 

The nationalization of agricultural land has yielded similar results.  During China's Great Leap Forward, villagers risked their lives to reallocate communal land into individual plots, a scheme starkly against the official Marxist ideology.  This informal system of property rights helped prevent starvation in the villages that adopted it.  Kate Xiao Zhou calls it How the Farmers Changed China.

 

Private agricultural plots emerged in the Soviet Union too, even with some government approval.  As Nellie Ohr explains in her Stanford dissertation “In 1930 Stalin had declared that [collective farm] members could own a few livestock and use a small private plot of land for personal consumption.  However, these rules were often unclear….”  Economist Gregory Grossman estimated that “the approximately 50 million such tiny ‘farms,’ whose area represent only about 3 percent of the national total of cultivated land, have a gross output which is more than one fourth the gross output of Soviet agriculture.”  [Paul Samuelson was infamously surprised that private agriculture in the USSR, cultivated by uneducated peasants without the benefit of scale economies, would be more productive than government farms.]  See also The Private Sector in Soviet Agriculture.

 

I recently visited some friends working at a major consulting firm in London.  Of course, the firm is a capitalist enterprise, but regarding its office space the management decided that the workspaces would be a common resource for all employees.  The management now has its own Spot moments, as employees use the same workspaces day after day, informally treating them as their own.

Economics Lessons from the Kibbutzim

In March 2023, I visited Ma'agan Michael Kibbutz and had a meeting with a former Israeli government official charged with managing the government bailout of about 200 kibbutzim.  Both offered empirical lessons about the challenges of communal living.

 

Background and Description

Ma'agan Michael was one of the few kibbutzim not bailed out.  It is one of the wealthier and more populous kibbutzim, with approximately 2000 residents. (Kibbutzim is the plural of kibbutz, which is described in more detail below but might be briefly described as a voluntary socialist community in Israel.)

 

In an earlier era, this kibbutz and the others followed a more egalitarian model.  For example, children were once raised communally in dormitories and would visit their biological parents for only three hours each day.  This alone was a big sacrifice for families for the sake of ideology, not to mention harmful actions that adults might occasionally take against children knowing that their parents were not around.

 

Now nuclear families live together in their own residence.  Every family receives an income that varies only according to the age and structure of their family.  As of 2023, that was about $3,500 per month.  Dining and laundry are in common. Rent, healthcare, formal education, and those two services are "free" -- not taken out of the $3,500.  A family can purchase and own household appliances for their residence.  Some get a washer and dryer because the community laundry takes too long.

 

Part of the kibbutz is something like a car rental agency.  They have a fleet of cars.  A central office holds the keys.  A member could pick up a car key and use a car for the day.  I believe that this was charged against their $3,500. Much of the accounting is done on computer with a fab system.

 

Some of the members are employed outside the kibbutz, but nonetheless must surrender all of their earnings to the community.  A partial exception would be earnings during periods of sabbatical.  One three-year sabbatical is permitted per lifetime.

 

Forty-five hours per week of work is required from all adults, with some categorical exceptions.  Mothers are required only 37.5 hours per week.  Retirees do not work at all; their Israeli-government pensions go to the community.

 

More than 90 percent of the land in Israel is government land, including the land allocated to Ma'agan Michael since at least the 1940s.  It is a prime seaside property, as you can see in this photo.  They once fished in the Mediterranean but now operate a couple of different kinds of fish farms on their property.  They also have a couple of factories for manufacturing.  The plastics factory they founded in 1963 has annual sales approaching $100 million, which helped them avoid the financial crises experienced by most other kibbutzim.

 

 

The common areas of the kibbutz can be compared to a college.  The dining room resembles the college cafeterias from my own college days, which are not as upscale as modern college dining facilities.  The grounds are less neatly kept than college campuses are (the beach photo is not typical in that regard).

 

Problems with communal living, even on a small scale

 

As Ran Abramitzky discusses in his book The Mystery of the Kibbutz, and labor economists have observed with employment contracts, work effort can be a problem when pay is not tied to performance.  The kibbutz, where members have often known each other since birth, tries to police this by watching each other.  They can expel a member from the kibbutz for poor performance, which is a serious punishment.

 

The kibbutz also looks hard at the work ethic of persons applying for membership.  However, membership in the kibbutz is voluntary, and individuals cannot be compelled to stay.  The most productive members have a significant financial incentive to leave.

 

Perhaps a more serious problem is occupational choice and human capital accumulation.  With uniform pay, members have little financial incentive to excel in their jobs, particularly in roles they find unfulfilling.  While the kibbutz covers the cost of formal education, much of human capital development in conventional labor markets occurs after formal schooling is completed.

 

Too much is democratized on the kibbutzim.  First, there is the challenge of reaching agreement, leading to significant and frustrating indecision.  Second, incentives are lacking to acquire and utilize information relevant to collective decision-making.  An individual could work hard to determine the right answer, but his vote hardly counts (the paradox of voting).

 

Financial decisions are a prime example.  Few kibbutz members have an incentive to learn about, say, present values.  About fifty years ago, numerous kibbutzim initiated projects that seemed viable if the time value of money were ignored, but were big losers from an NPV perspective.  The abundance of failures among those projects was a major reason that kibbutzim would later be bailed out.

 

Equality, including gender equality, was an important ideal of the early kibbutzim and, to a large extent, still today.  Nevertheless, despite the ideology, gender segregation by occupation seems just as strong as it is off kibbutz.  All but one of the laundry employees I saw were female.  All of the tractors and heavy machinery were operated by men.