In this message I will discuss some horrible ideas, so let’s start with good news!
I have Previously studied Tech deducts. According to my most important data source fired.fyiThe weekly number of dismissals has been steadily decreased. In the US this is compensated by a large number of fired in the government sector, but the trend for technical employees seems good. However, the number of dismissals remains considerably higher than its low in 2022, which means that technical employees have less job security than before.
In this message I want to consider two possible mechanisms that can motivate companies to perform mass dismissals.
You may notice that there is a negative autocorrelation in the data. In other words, weeks with a large number of dismissals are followed by weeks with a low number of dismissals. This applies to the technical sector as a whole, but it can also be true within individual companies.
A good example is the case of Brittany Pietsch Who filmed himself in 2024 was fired from Cloudflare. An interesting aspect of the Pietsch case was that HR was unable to give her a reason for her termination, because her performance was considered satisfactory. Taking this on nominal value leads to a question: why would a company hire someone and then dismiss a few months later? What’s in it for them?
The principal of shareholder value, also known as the Friedman Doctrine After Milton Friedman1states that the aim of each company is to maximize the profit of its shareholders. This controversial idea has become a dogma in the business world over the past 40 years and has led to higher management being largely paid in stock and (especially) stock options. The idea is that the task of the CEO is to increase the share price, whatever happens. One way to motivate the CEO to do this is to link the total reimbursement to the stock price. Did the stock price not risen this quarter? Difficult happiness, your options are worthless!
Traditionally, dismissal is supposed to print the share price, because they indicate that a company is not doing well. But nowadays this no longer seems to be the case. An announcement of dismissals could increase The stock price. In the business world, Income per employee Become an ever -popular statistics to compare companies with each other. The less the employees, the larger the number.
It follows that it can be logical for a company to explain people, even if there is a business reason to. But you have to hire people before you can cover them. So, hire people. This is good for your stock price in the long term because your company seems to be growing. Then put a massive dismissal. This is good for your share price in the short term because it increases your income per employee. Rinse, repeat, collect your money!
I call this strategy To wash Because it is analogous to the situation on the stock market where you sell a share to yourself to make it look valuable. This idea was used by famous Stockmanipulators such as Daniel Drew. Nowadays it is illegal, but it is still being washed when people invent a new financial market. For example, when NFTs were popular, people earned money through Sell NFTs to themselves To encourage other people to buy them for a bloated price.
Daniel Drew, Wikipedia
Do the modern CEOs of the company that do the same with human lives? Possible. Some companies practice Ranking (or Rank As Jack Welch2 called it) in which the lower x percent of employees are sacrificed in a regular round of dismissals. Was hiring are more or less the same only without the ranking part.
If companies really do this, is there a problem with it? It has the potential to cause a lot of economic damage. If people are afraid of random fired, they will probably increase their precautions and not be so productive (because they have to spend time to ensure that they are prepared for a search for a job.) As Ha-Joon Chang explains in 23 things they don’t tell you about capitalismThing 10, there are hidden costs for work uncertainty that usually do not appear in economic models3.
How would we know if CEOs hire people pure to let someone shoot? Well, we could ask them! It appears that the CEOs of Corporate America are more than happy to talk about their mass dismissal plans. But According to the Wall Street JournalThe reason why many CEOs want to abolish people is not to increase their share price, but to replace people with AI.
So what happens if you replace people with AI? Can this also lead to a cycle of recruitment and shooting? I have done some modeling to find out and it seems that the answer is to a certain extent yes.
A capital allocation model
Suppose a company has a fixed share of $ K $ in capital and produces output according to the COBB-Douglas production function4
\[Y = A K^\alpha L^{1-\alpha}.\]
Suppose the company is a price contractor on the labor market (ie employees can be hired at a fixed wage $ W $) and the company wants to maximize the profit. Profit is $ Y-WL $ and so the company will hire a number of work
\[L^* = K((1-\alpha)A/w)^{1/\alpha}.\]
Suppose someone invents an AI and the company has the opportunity to hire AI work instead. The AI requires a wage of $ 0 $, so the company will of course replace all human labor by AI. But as data scientists, we know that AI models will break down over time as a result of data deviation. Therefore, assume that the AI Labor $ u_t $ meets a comparison of the form5
\[U_{t+1} = (1-\delta)U_t + bL_t\]
Where $ (1- \ Delta) $ represents the depreciation of the AI quality over time, $ B $ represents the amount of new training data generated by employees at time $ t $, and $ u_t $ is the amount of equivalent can do the AI $ $.
The total win at time is $ t $ is
\[\Pi_t = AR_t^\alpha U_t^{1-\alpha} + A(K-R_t)^\alpha L_t^{1-\alpha} – wL_t\]
Where $ R_T $ is the capital that is assigned to the AI employees (remember that total capital $ K $ has been determined.) We can calculate the choices of $ R_T $ and $ l_t $ that maximize the profit by Calculus. This leads to two possible solutions. Either the company devotes all its capital to AI employees and $ R_T = K, L_T = 0 $ or otherwise
\[\begin{aligned}
L_t &= (A(1-\alpha)/w)^{1/\alpha}(K-R_t) \\
R_t &= (w/A(1-\alpha))^{1/\alpha}U_t
\end{aligned}\]
It follows that $ \ frac {\ partial l_t} {\ partially l_ {t-1}} <0 $. This is how you can get fluctuations in $ l_t $, or cycles of recruitment and shooting.
Here a numerical example has been calculated in R:
N <- 50 # number of time periods
R <- U <- L <- Y <- Pi <- rep(0, N)
alpha <- 1/3
K <- 1
w <- 0.1
A <- 1
delta <- 0.8
b <- 1
# initial values
L0 <- ((1-alpha)/w)^(1/alpha)
L[1] <- L0
U[1] <- 0 # start with no AI at time 1
for (i in 2:N){
U[i] <- (1-delta)*U[i-1] + b*L[i-1]
Rn <- (w/(1-alpha))^(1/alpha) * U[i]
if (Rn > K){
R[i] <- K
L[i] <- 0
} else {
R[i] <- Rn
L[i] <- ((1-alpha)/w)^(1/alpha) * (K-R[i])
}
Y[i] <- R[i]^(alpha)*U[i]^(1-alpha) + (K-R[i])^(alpha)*L[i]^(1-alpha)
Pi[i] <- Y[i] - w*L[i]
}
Initially (on time t=2 Because rexes vectors starting at 1), everyone is fired because he is being replaced by AI. But because the implementation of the AI breaks down over time, some people must be hired and then fired again than … In the end, a balance is achieved in which less human work is required, but the cycles of hiring and shooting no longer happen.
Could such a thing happen in the real world? Well, there is some proof for it! You may remember that buy-now-pay-company Klarna Recently announced That it repeated employees after they had fired them because of AI. Just like in the model, the AI who had to replace the employees turned out not to be everything it was to be!
So there you are; Two possible reasons why you would expect you to expect cycles from recruitment and shoot at the labor market. The idea of Washeren is attractive, but it is only logical for the public traded for the public6. The idea of AI replacement is certainly realistic and also frightening.
For example, some economists have articles about the possible impact of AI on the labor market This article by Pascal Stiefenhofer. But the research tends to concentrate on where we will end, not so much how we will get there in the short term.
It seems that it can be a bumpy ride!
1: The following quote comes from A YouTube interview with Joseph Stiglitz
I wrote a paper in which it was explained why companies that maximize their stock market would not lead to the general well -being of society. […] Around the same time, Milton Friedman wrote a very influential article that was published in the magazine New York Times about why companies should have to Maximize their market value. I was invited to the University of Chicago to give a lecture about my newspaper. […] After the seminar we sat outside the seminar room and he said, “Joe. You were wrong.” I said, “Spleen. Tell me where I made a mistake in my analysis.” He said: “Joe. You are wrong. I know that the maximization of the shareholders is optimal. It’s good for society”. I said: “I worked very hard to show that you are wrong. Just explain where I made a mistake in my analysis, “and he said,” Joe. You are wrong. “And it was very clear at that time that he had become an ideal.
2: Welch, as the boss of General Electric, was one of the most controversial and most successful exponents of shareholder value that he later, in an interview in 2009 with the Financial timescalled “The stupidest idea in the world.”
3: A 1979 Model of Jovanovic Consider how much effort an employee must spend on his search for work versus his current job. It is clear that if you expect to be fired at any time without reason, you might as well make minimal efforts on your current job and spend all your time learning skills that can help you in your next job.
4: The COBB-Douglas function may not be a suitable model for an individual company, but the conclusion would be similar to any reasonable choice of production function.
5: Again, the exact comparison is not important. We only require that the quality of the AI decreases over time if there is no human work to maintain it or to generate new training data for it.
6: My personal experience has been the opposite. The last time I had the feeling that I had a reasonably safe job was when I worked for a listed company, while later jobs in private companies have never felt particularly safe. But this can simply be due to the general economic conditions.
Related
#replacement #wash #RBloggers