Connect with us

Uncategorized

COVID-19 in Latin America and the Caribbean: A Data Review

COVID-19 in Latin America and the Caribbean: A Data Review

Published

on

Though COVID-19 was slow to appear in worrying numbers in Latin America and the Caribbean, the number of infected people in the region is now rising rapidly. Between March 1 and April 1, total confirmed cases of the novel coronavirus in Latin America rose from 5 to 25,500. In the first half of April, that number nearly tripled to 75,200. The US currently has the highest rate of reported COVID-19 cases per capita of any country in the Western Hemisphere, but this may be primarily because it saw its first case weeks earlier. Latin America may still have a similarly overwhelming case load coming its way.

Like in other regions around the world, the impacts of the pandemic are not evenly distributed throughout Latin America and the Caribbean. Infections are concentrated in certain countries and certain areas within countries; policy responses and health care systems differ from country to country, and from region to region within countries. 

Current Cases

The majority of confirmed Latin American cases are in just a handful of countries, with Brazil making up more than a third of all reported cases. Looking at the raw number of reported cases can help give us an idea of where the disease is, but looking only at the overall numbers may obscure serious outbreaks in smaller nations. Additionally, Figure 1 makes it clear that some countries have large numbers of total cases per capita, but much smaller numbers of active cases per capita due to recoveries and deaths. Figure 2 presents national data of confirmed active COVID-19 cases per 1,000 people. 

As can be seen, measuring confirmed cases in this manner highlights some potentially significant outbreaks that have thus far largely occurred under the radar. The highest overall active rate of infection right now can be found in Panama, which has more than double the rate seen in Ecuador, the Dominican Republic, and Chile, the next most-infected nations. Along with Peru, the outbreaks in Ecuador and Chile suggest that COVID-19 is more widespread on the Western Pacific coast of South America. Accounting for population size also illustrates that there appears to be significant outbreaks in some smaller Caribbean nations as well, with the Dominican Republic, Saint Kitts and Nevis, Antigua and Barbuda, Barbados, and Grenada all making the top ten.

There are reasons to question these numbers. First, even if they all reported in good faith, each nation has its own sets of standards and collection mechanisms for reporting this data, meaning that each nation has differences in how their COVID-19 numbers are put together. Even more problematic, however, is the matter of testing. COVID-19 cases, recoveries, and deaths can only be reported if the virus is first diagnosed through testing, and the availability of testing varies widely. The inconsistency in testing across the region may also prevent us from having an accurate picture of the disease’s spread. Though not all countries in Latin America are reporting their testing numbers, a look at the data reveals that many nations in the region may be drastically underreporting confirmed cases. 

To get an idea of which countries are testing more than others, Figure 3 presents tests per 1,000 people for those countries with available data, with darker green illustrating more testing and white illustrating less.

The gaps between testing rates and reported active case rates can shed some light on who is underreporting most. Figure 4 plots tests per million people against active COVID-19 cases with a trendline that indicates how much an additional amount of testing yields in additional cases on average. Venezuela is excluded from this chart because they have recently received large shipments of medical supplies from China which dramatically expanded their testing capacity in just the last week, while their case numbers have remained relatively low; this suggests that it’s unlikely they’ve had a chance to deploy the tests and report back more accurate numbers accordingly.

If a country appears above the trendline, it indicates that it has a higher number of active cases than we would expect for a country in the region given its number of tests; if it appears below, it has fewer cases than would be expected.

This visualization tells us that the high levels of COVID-19 reported in some countries, like Chile and Panama, is in part due to the fact that they have some of the highest testing rates in Latin America. Unlike Chile, however, Panama’s placement high above the trendline indicates that it does in fact have a very serious outbreak on hand. Other countries with above-average cases per test include Antigua and Barbuda, the Dominican Republic, Grenada, Ecuador, and St. Vincent and the Grenadines. Of the countries that have reported their testing rates, Guatemala, Haiti, Honduras, and Bolivia have the lowest. It is no coincidence that these are some of the poorest and most underdeveloped countries in Latin America: testing rates correlate significantly with both GDP per capita and the UN’s Inequality-adjusted Human Development Index. Because the accuracy of coronavirus rates rises with more testing, we can’t be sure how widespread COVID-19 actually is in these countries.

There is another complication: the first cases of COVID-19 appeared at different times in different countries, and there is a clear relationship between how long it’s been since the beginning of an outbreak and how many total cases a nation has had. As such, it’s worth keeping an eye on the last countries to see cases appear, as these are nations likely to witness the most future growth. Examples include Belize, El Salvador, Grenada, Haiti, and Nicaragua.

On a more positive note, a few nations like Costa Rica, Cuba, and Uruguay appear to have handled the crisis fairly well so far, with a low number of active cases per thousand even when taking into account relatively high testing rates and the fact that their first cases appeared over a month ago.

Health Care Systems

Like the reported numbers of COVID-19 itself, health care systems and policies vary significantly across Latin America. Data on these types of metrics can help us get an idea of which nations are most ready to contain and treat coronavirus, and which may be least prepared to do so. Two simple measures of health care system capacity are hospital beds per capita and doctors and nurses per capita, serving as proxies for both basic health infrastructure and the quantity of medical personnel. In Figure 5, these two measures are plotted against one another for comparison, with each nation shaded darker blue the wealthier it is on a per capita basis.

For the most part, wealthier nations in Latin America have both more hospital beds and health professionals per capita than their poorer neighbors, and will thus be able to handle a larger strain on their health care system. Some countries stand out. Argentina, Brazil, Chile, and Barbados have all clearly put major effort into developing certain aspects of their health care system, while Cuba’s system is by far the most prepared in terms of sheer capacity despite Cuba being less wealthy. Panama, the worst-off country in the region right now, has put significantly less investment in health care capacity than other nations of a similar relatively high-income. Dominica and St. Vincent and the Grenadines punch well-above their weight here considering their GDP per capita. Finally, three nations stand out in their unique unpreparedness for a serious health crisis: Guatemala, Haiti, and Honduras. This is especially important to note as the data presented above indicate that these three countries are all among the countries with the lowest testing rates. Not only are the countries particularly vulnerable, but we also have little information about how widespread the virus is within them.

It perhaps comes as little surprise that three of the most vulnerable countries are also three of the most impoverished nations in the hemisphere. Just as people living below the poverty line are most vulnerable, so too are the most impoverished countries. While the countries with the largest outbreaks in raw numbers will likely receive the lion’s share of attention, it is important to keep in mind that small states — including many small island nations — will face significant outbreaks on their own right.

The post COVID-19 in Latin America and the Caribbean: A Data Review appeared first on Center for Economic and Policy Research.

Read More

Continue Reading

Uncategorized

February Employment Situation

By Paul Gomme and Peter Rupert The establishment data from the BLS showed a 275,000 increase in payroll employment for February, outpacing the 230,000…

Published

on

By Paul Gomme and Peter Rupert

The establishment data from the BLS showed a 275,000 increase in payroll employment for February, outpacing the 230,000 average over the previous 12 months. The payroll data for January and December were revised down by a total of 167,000. The private sector added 223,000 new jobs, the largest gain since May of last year.

Temporary help services employment continues a steep decline after a sharp post-pandemic rise.

Average hours of work increased from 34.2 to 34.3. The increase, along with the 223,000 private employment increase led to a hefty increase in total hours of 5.6% at an annualized rate, also the largest increase since May of last year.

The establishment report, once again, beat “expectations;” the WSJ survey of economists was 198,000. Other than the downward revisions, mentioned above, another bit of negative news was a smallish increase in wage growth, from $34.52 to $34.57.

The household survey shows that the labor force increased 150,000, a drop in employment of 184,000 and an increase in the number of unemployed persons of 334,000. The labor force participation rate held steady at 62.5, the employment to population ratio decreased from 60.2 to 60.1 and the unemployment rate increased from 3.66 to 3.86. Remember that the unemployment rate is the number of unemployed relative to the labor force (the number employed plus the number unemployed). Consequently, the unemployment rate can go up if the number of unemployed rises holding fixed the labor force, or if the labor force shrinks holding the number unemployed unchanged. An increase in the unemployment rate is not necessarily a bad thing: it may reflect a strong labor market drawing “marginally attached” individuals from outside the labor force. Indeed, there was a 96,000 decline in those workers.

Earlier in the week, the BLS announced JOLTS (Job Openings and Labor Turnover Survey) data for January. There isn’t much to report here as the job openings changed little at 8.9 million, the number of hires and total separations were little changed at 5.7 million and 5.3 million, respectively.

As has been the case for the last couple of years, the number of job openings remains higher than the number of unemployed persons.

Also earlier in the week the BLS announced that productivity increased 3.2% in the 4th quarter with output rising 3.5% and hours of work rising 0.3%.

The bottom line is that the labor market continues its surprisingly (to some) strong performance, once again proving stronger than many had expected. This strength makes it difficult to justify any interest rate cuts soon, particularly given the recent inflation spike.

Read More

Continue Reading

Uncategorized

Mortgage rates fall as labor market normalizes

Jobless claims show an expanding economy. We will only be in a recession once jobless claims exceed 323,000 on a four-week moving average.

Published

on

Everyone was waiting to see if this week’s jobs report would send mortgage rates higher, which is what happened last month. Instead, the 10-year yield had a muted response after the headline number beat estimates, but we have negative job revisions from previous months. The Federal Reserve’s fear of wage growth spiraling out of control hasn’t materialized for over two years now and the unemployment rate ticked up to 3.9%. For now, we can say the labor market isn’t tight anymore, but it’s also not breaking.

The key labor data line in this expansion is the weekly jobless claims report. Jobless claims show an expanding economy that has not lost jobs yet. We will only be in a recession once jobless claims exceed 323,000 on a four-week moving average.

From the Fed: In the week ended March 2, initial claims for unemployment insurance benefits were flat, at 217,000. The four-week moving average declined slightly by 750, to 212,250


Below is an explanation of how we got here with the labor market, which all started during COVID-19.

1. I wrote the COVID-19 recovery model on April 7, 2020, and retired it on Dec. 9, 2020. By that time, the upfront recovery phase was done, and I needed to model out when we would get the jobs lost back.

2. Early in the labor market recovery, when we saw weaker job reports, I doubled and tripled down on my assertion that job openings would get to 10 million in this recovery. Job openings rose as high as to 12 million and are currently over 9 million. Even with the massive miss on a job report in May 2021, I didn’t waver.

Currently, the jobs openings, quit percentage and hires data are below pre-COVID-19 levels, which means the labor market isn’t as tight as it once was, and this is why the employment cost index has been slowing data to move along the quits percentage.  

2-US_Job_Quits_Rate-1-2

3. I wrote that we should get back all the jobs lost to COVID-19 by September of 2022. At the time this would be a speedy labor market recovery, and it happened on schedule, too

Total employment data

4. This is the key one for right now: If COVID-19 hadn’t happened, we would have between 157 million and 159 million jobs today, which would have been in line with the job growth rate in February 2020. Today, we are at 157,808,000. This is important because job growth should be cooling down now. We are more in line with where the labor market should be when averaging 140K-165K monthly. So for now, the fact that we aren’t trending between 140K-165K means we still have a bit more recovery kick left before we get down to those levels. 




From BLS: Total nonfarm payroll employment rose by 275,000 in February, and the unemployment rate increased to 3.9 percent, the U.S. Bureau of Labor Statistics reported today. Job gains occurred in health care, in government, in food services and drinking places, in social assistance, and in transportation and warehousing.

Here are the jobs that were created and lost in the previous month:

IMG_5092

In this jobs report, the unemployment rate for education levels looks like this:

  • Less than a high school diploma: 6.1%
  • High school graduate and no college: 4.2%
  • Some college or associate degree: 3.1%
  • Bachelor’s degree or higher: 2.2%
IMG_5093_320f22

Today’s report has continued the trend of the labor data beating my expectations, only because I am looking for the jobs data to slow down to a level of 140K-165K, which hasn’t happened yet. I wouldn’t categorize the labor market as being tight anymore because of the quits ratio and the hires data in the job openings report. This also shows itself in the employment cost index as well. These are key data lines for the Fed and the reason we are going to see three rate cuts this year.

Read More

Continue Reading

Uncategorized

Inside The Most Ridiculous Jobs Report In History: Record 1.2 Million Immigrant Jobs Added In One Month

Inside The Most Ridiculous Jobs Report In History: Record 1.2 Million Immigrant Jobs Added In One Month

Last month we though that the January…

Published

on

Inside The Most Ridiculous Jobs Report In History: Record 1.2 Million Immigrant Jobs Added In One Month

Last month we though that the January jobs report was the "most ridiculous in recent history" but, boy, were we wrong because this morning the Biden department of goalseeked propaganda (aka BLS) published the February jobs report, and holy crap was that something else. Even Goebbels would blush. 

What happened? Let's take a closer look.

On the surface, it was (almost) another blockbuster jobs report, certainly one which nobody expected, or rather just one bank out of 76 expected. Starting at the top, the BLS reported that in February the US unexpectedly added 275K jobs, with just one research analyst (from Dai-Ichi Research) expecting a higher number.

Some context: after last month's record 4-sigma beat, today's print was "only" 3 sigma higher than estimates. Needless to say, two multiple sigma beats in a row used to only happen in the USSR... and now in the US, apparently.

Before we go any further, a quick note on what last month we said was "the most ridiculous jobs report in recent history": it appears the BLS read our comments and decided to stop beclowing itself. It did that by slashing last month's ridiculous print by over a third, and revising what was originally reported as a massive 353K beat to just 229K,  a 124K revision, which was the biggest one-month negative revision in two years!

Of course, that does not mean that this month's jobs print won't be revised lower: it will be, and not just that month but every other month until the November election because that's the only tool left in the Biden admin's box: pretend the economic and jobs are strong, then revise them sharply lower the next month, something we pointed out first last summer and which has not failed to disappoint once.

To be fair, not every aspect of the jobs report was stellar (after all, the BLS had to give it some vague credibility). Take the unemployment rate, after flatlining between 3.4% and 3.8% for two years - and thus denying expectations from Sahm's Rule that a recession may have already started - in February the unemployment rate unexpectedly jumped to 3.9%, the highest since February 2022 (with Black unemployment spiking by 0.3% to 5.6%, an indicator which the Biden admin will quickly slam as widespread economic racism or something).

And then there were average hourly earnings, which after surging 0.6% MoM in January (since revised to 0.5%) and spooking markets that wage growth is so hot, the Fed will have no choice but to delay cuts, in February the number tumbled to just 0.1%, the lowest in two years...

... for one simple reason: last month's average wage surge had nothing to do with actual wages, and everything to do with the BLS estimate of hours worked (which is the denominator in the average wage calculation) which last month tumbled to just 34.1 (we were led to believe) the lowest since the covid pandemic...

... but has since been revised higher while the February print rose even more, to 34.3, hence why the latest average wage data was once again a product not of wages going up, but of how long Americans worked in any weekly period, in this case higher from 34.1 to 34.3, an increase which has a major impact on the average calculation.

While the above data points were examples of some latent weakness in the latest report, perhaps meant to give it a sheen of veracity, it was everything else in the report that was a problem starting with the BLS's latest choice of seasonal adjustments (after last month's wholesale revision), which have gone from merely laughable to full clownshow, as the following comparison between the monthly change in BLS and ADP payrolls shows. The trend is clear: the Biden admin numbers are now clearly rising even as the impartial ADP (which directly logs employment numbers at the company level and is far more accurate), shows an accelerating slowdown.

But it's more than just the Biden admin hanging its "success" on seasonal adjustments: when one digs deeper inside the jobs report, all sorts of ugly things emerge... such as the growing unprecedented divergence between the Establishment (payrolls) survey and much more accurate Household (actual employment) survey. To wit, while in January the BLS claims 275K payrolls were added, the Household survey found that the number of actually employed workers dropped for the third straight month (and 4 in the past 5), this time by 184K (from 161.152K to 160.968K).

This means that while the Payrolls series hits new all time highs every month since December 2020 (when according to the BLS the US had its last month of payrolls losses), the level of Employment has not budged in the past year. Worse, as shown in the chart below, such a gaping divergence has opened between the two series in the past 4 years, that the number of Employed workers would need to soar by 9 million (!) to catch up to what Payrolls claims is the employment situation.

There's more: shifting from a quantitative to a qualitative assessment, reveals just how ugly the composition of "new jobs" has been. Consider this: the BLS reports that in February 2024, the US had 132.9 million full-time jobs and 27.9 million part-time jobs. Well, that's great... until you look back one year and find that in February 2023 the US had 133.2 million full-time jobs, or more than it does one year later! And yes, all the job growth since then has been in part-time jobs, which have increased by 921K since February 2023 (from 27.020 million to 27.941 million).

Here is a summary of the labor composition in the past year: all the new jobs have been part-time jobs!

But wait there's even more, because now that the primary season is over and we enter the heart of election season and political talking points will be thrown around left and right, especially in the context of the immigration crisis created intentionally by the Biden administration which is hoping to import millions of new Democratic voters (maybe the US can hold the presidential election in Honduras or Guatemala, after all it is their citizens that will be illegally casting the key votes in November), what we find is that in February, the number of native-born workers tumbled again, sliding by a massive 560K to just 129.807 million. Add to this the December data, and we get a near-record 2.4 million plunge in native-born workers in just the past 3 months (only the covid crash was worse)!

The offset? A record 1.2 million foreign-born (read immigrants, both legal and illegal but mostly illegal) workers added in February!

Said otherwise, not only has all job creation in the past 6 years has been exclusively for foreign-born workers...

Source: St Louis Fed FRED Native Born and Foreign Born

... but there has been zero job-creation for native born workers since June 2018!

This is a huge issue - especially at a time of an illegal alien flood at the southwest border...

... and is about to become a huge political scandal, because once the inevitable recession finally hits, there will be millions of furious unemployed Americans demanding a more accurate explanation for what happened - i.e., the illegal immigration floodgates that were opened by the Biden admin.

Which is also why Biden's handlers will do everything in their power to insure there is no official recession before November... and why after the election is over, all economic hell will finally break loose. Until then, however, expect the jobs numbers to get even more ridiculous.

Tyler Durden Fri, 03/08/2024 - 13:30

Read More

Continue Reading

Trending