Beat the Press is Dean Baker's commentary on economic reporting. Dean Baker is co-director of the Center for Economic and Policy Research (CEPR).

Follow on Twitter Like on Facebook Subscribe by E-mail RSS Feed

I have a hot tip for Washington Post reporters: politicians aren't always honest. As a result when they say they believe something, it doesn't mean they really believe it.

This means that the Washington Post likely misled its readers in a discussion of Republican tax cut proposals when it told them:

"Republicans believe the corporate rate cut and other incentives will stimulate economic growth, offsetting the revenue loss."

The reality is the Washington Post's reporters have no clue what the Republicans pushing tax cuts really believe about the impact of tax cuts on growth. If these Republicans were at all familiar with with the evidence, they would not expect their tax cuts to have much, if any, positive impact on growth.

So while it is possible that Republicans believe in something that is not true, it is possible that they are deliberately deceiving the public. It is also possible that they have no clue whatsoever about the impact of tax cuts on the economy, just as they have no clue about the impact of their health care proposals. They are simply voting as their funders are telling them to.

Newspapers are supposed to report the facts, not make them up, as the Post is doing here.

Add a comment

With the Republicans promising big tax cuts for their wealthy backers, we are again hearing talk about the budget deficit and national debt. Needless to say, most is pretty badly confused.

At the most basic level, insofar as there is a burden of the debt, it is the interest payments on the debt. This is the amount of money that the government has to cough up each year to pay bondholders as opposed to using for other purposes.

While the debt is high relative to GDP, interest on the debt is actually fairly low. It is currently around 0.8 percent of GDP, or roughly $160 billion a year. This is near a post-World War II low. In the 1990s, the interest peaked at more than 3.0 percent of GDP.

Many people overstate the interest burden because they ignore the money that the Federal Reserve Board refunds to the government, which is presently around $90 billion a year. The Fed is collecting a substantial portion of the interest paid by the federal government due to the fact that it holds a large amount of government bonds. It keeps some of the interest to fund its operations and then rebates the rest to the Treasury.

Incredibly, there has been literally no coverage in major news outlets of the budgetary implications of the Fed's plans to reduce its asset holdings. When the Fed sells off these assets, the interest will instead be paid out to the people who buy the bonds rather than refunded to the Treasury. The difference could come to as much as $600 billion over the next decade, roughly the amount of money at stake with Obamacare repeal, but no one seems to care.

Add a comment

The Washington Post misled readers in its discussion of Republican claims that its tax cuts will lead to a large boost to GDP growth. The piece quotes Kent Smetters, an economist at the University of Pennsylvania, as saying the Republican growth projections did not take account of debt. This is wrong.

The real issue is whether the projections take account of how close the U.S. economy is to its potential level of output. If the current level of demand is near the point where the economy is hitting serious supply constraints, then the Republican tax cuts will not have much impact on growth. On the other hand, if there is still considerable excess capacity in the form of unemployed and underemployed workers, then it may be possible to increase growth by increasing demand, such as the tax cuts.

Debt is a meaningless concept in this context. Debt would only matter insofar as the flow of income in the form of interest payments on bonds create a source of demand that pull resources away from other uses. With interest payments near a post-war low as a share of GDP, this should not be a major issue for the foreseeable future.

Also, direct debt is only one way in which the government commits flows of future income. Government-granted patent and copyright monopolies are actually much more important in determining future flows of income than debt. In the case of prescription drugs alone, patent and related protections raise the price of drugs by close to $370 billion a year over the free market price, a bit less than 2.0 percent of GDP. This is considerably larger than the current interest burden of the debt, which is approximately 1.6 percent of GDP, net of money refunded from the Federal Reserve Board to the Treasury.

These monopolies are effectively like privately collected taxes. The government grants them as a way to pay for research and creative work. If anyone were really concerned about the burden created by government debt, they would factor in the cost of these monopolies to the public. The decision not to include the costs from patent and copyright monopolies in assessments of the debt makes for a fundamentally dishonest discussion of the issue.

Add a comment

Paul Krugman does a nice job dissecting the logic, or lack thereof, of Republican efforts to dismantle Obamacare. He points out that one of the games played by the Republicans is claiming that the Graham-Cassidy bill would increase Medicaid spending. It makes this claim based on the fact that nominal Medicaid spending would increase under the bill. While this is true, the bill would hugely cut spending compared with the baseline which factors in both projected increases in the number of people covered and the medical cost inflation.

It might be harder to get away with this cheap trick if papers like the New York Times used budget reporting to actually inform readers. As it is, budget reporting usually doesn't put numbers in any context, which makes them absolutely meaningless to the vast majority of readers. This point was acknowledged a few years back by both Margaret Sullivan and David Leonhardt, who were at the time the NYT's public editor and Washington editor. 

In spite of the acknowledgment that its budget reporting does nothing to inform the vast majority of its readers, the paper has done nothing to change the practice. (Does the paper really find it hard to find reporters who know basic arithmetic?) As a result, it is much easier for Republicans to lie when they want to do something like massively cutting back Medicaid spending.

Add a comment

It is more than a bit bizarre that no one seems to pay any attention to the budgetary implications of the Fed's decision to start selling off its assets. The impact is potentially fairly large in the scheme of things, possibly as much as $600 billion over the next decade. This is equal to roughly 0.5 percent of GDP. It's pretty much the same number at stake in the various Obamacare repeal efforts.

For some reason, none, as in absolutely zero, of the news stories I have seen or heard about the asset sales mentioned its impact on the budget. It was the same story back in June when the Fed raised the issue at its meeting that month.

It's a bit hard to understand how reporters at the New York Times, Washington Post, NPR, and elsewhere can independently decide that adding hundreds of billions of dollars to the budget deficit over the next decade is not worth mentioning. (The basic story can be found here.)

Add a comment

Germany's official measure of unemployment is constructed differently than the U.S. measure. They count people working part-time who want full-time jobs as being unemployed. In contrast, these people are counted in the United States as being employed. As a result, the official measure is not directly comparable to the U.S. measure. Fortunately, the OECD constructs a "harmonized" unemployment rate which essentially applies the U.S. methodology to the unemployment measures for other countries.

Since the OECD measure is readily available, it is difficult to see why the NYT used the German official measure in an article on growing poverty and inequality in Germany. The piece tells readers that Germany's unemployment rate is 5.7 percent. This is the official German measure. The unemployment rate using the OECD's harmonized measure is 3.7 percent. Since few readers are likely to be familiar with the methodology used to construct the German measure, using the 5.7 percent figure is misleading to them. 

Add a comment

The NYT criticized the Republican tax-cutting plans. In particular, it focuses on the plan to apply a 15 percent or 25 percent corporate tax rate for all business pass-through income. This is in place of the current individual income tax rate, which could be as high as 39.6 percent.

The editorial argues that this would be a huge tax break for partners in hedge funds and real estate developers (like Donald Trump), who typically get their income through pass-through corporations. While this is true, it is only part of the story.

Allowing owners of pass-through corporations to just pay the corporate tax rate instead of the individual income tax rate would be a huge loophole that every higher paid person would rush to take advantage of. For example, a highly paid medical specialist would reorganize their business as a pass-through corporation to pay a lower tax rate. The same would be true of other professionals. 

In effect, this would allow the vast majority of high-income individuals to pay a lower tax rate than school teachers or firefighters while creating an enormous tax shelter industry. This is truly awful from the standpoint of policy (increasing inequality while reducing efficiency), but if the point is to give more money to the rich, it gets the job done. 

Add a comment

Bryce Covert has an interesting column in the NYT arguing that Equifax and the other two private credit agencies be replaced with a public system. There does seem to be a good case here.

After all, what do we get from competition in this story? As Covert points out, the credit agencies don't work for consumers, they work for the people who buy the data. This means that they don't really have much incentive to ensure their information about us is accurate and to make sure their systems are not hacked.

What is difficult understand is how current laws allow these agencies to be profitable. Suppose these credit agencies were held liable for their mistakes. If a person is denied a job or is unable to buy a home because of an erroneous credit report, this would seem to warrant tens of thousands of dollars in damages.

Covert tells us that one in four credit reports contain a major error. Suppose that leads to 100,000 people a year to suffer serious consequences, that would be less than 0.25 percent of the people with serious errors in their report. If the average damages come to $15,000 (including legal fees), this would run to $1.5 billion in annual damage payments. If 1.0 percent of the people with erroneous reports suffered consequences, it would come to $6 billion a year.

Add a comment

It's good to see the NYT has diverse voices on economic issues. In his "Economic Scene" column on the economy's supposed labor shortage, Porter argued that "raising barriers to imports — inviting retaliation from trading partners — is exactly the wrong approach." Three days ago, the NYT had an editorial arguing strongly for increased protectionism in the form of stronger and longer copyright and patent protections. These forms of protection not only raise prices and slow growth, but they redistribute income upward, taking money from most of the population and giving it to those who stand to gain from patent and copyright protection.

It is worth noting that the assertion that the economy is now or will in the near future be suffering from a labor shortage seems dubious. Employment rates for prime-age men at all levels of educational attainment are still below their 2000 rate. That seems hard to reconcile with the view that we have a labor shortage. In the same vein, the percentage of unemployment due to people voluntarily quitting their jobs is still relatively low, suggesting that workers do not feel very secure about their labor market prospects. And, workers are still far from recovering their share of output, with the profit share still well above the pre-recession level.

For these reasons, we should be skeptical about the existence of a labor shortage. It is also worth noting that a labor shortage is 180 degrees at odds with the often asserted problem that the robots will take all the jobs.

Add a comment

After screaming about budget deficits throughout the Obama administration, Republicans in Congress are apparently planning to pass tax cuts that will substantially increase the budget deficit from the baseline projections. The NYT decided to help them in this effort by printing without comment their absurd claims about growth.

According to the NYT, Senator Ron Johnson, a member of the Budget Committee, said:

"Just going from 2 to 3 percent growth adds about $14 trillion of economic activity over a decade, $2 to $3 trillion of revenue to the federal government."

This would be like saying that we would have a great baseball team if we just brought Babe Ruth, Lou Gehrig, and Jackie Robinson back to life. Presumably, the NYT would recognize that anyone who said that about baseball was either lying or seriously out of touch with reality and would point this fact out to their readers.

Similarly, the idea that we have a simple route to "just" raise the rate of annual growth from 2 to 3 percent, is equally absurd, but the NYT treated it as though it is a comment that a sane person could have honestly said. That's great stenography, but truly awful reporting.

Add a comment

Glenn Kessler, the Washington Post's Fact Checker, is trying to be even-handed in assessing the claim by advocates of single-payer health insurance of large potential administrative savings from switching to a universal Medicare-type system. Unfortunately, he gives too much credence to an insurance industry funded report (identified as such in the piece), which whittles away at the difference.

The basic story is that Medicare's administrative costs, as shown in the Medicare Trustees Report come to less than 2.0 percent of annual benefits. (Table II.B1 puts administrative costs for 2016 at $9.2 billion, with total payments at $669.5 billion.) The Centers for Medicare and Medicaid Services puts the administrative costs for private insurance at $216.3 billion for 2016 (Table 2), with total spending of $1,135.4 billion (Table 3) for a ratio administrative expenses as a percent of related outlays of 23.5 percent. (The administrative expenses are deducted from total spending to get health related outlays.)

While there is some room for fudging both of these numbers, it is pretty hard to make the difference go away. The industry funded study ups the Medicare costs to 5.2 percent by allocating a large amount of non-Medicare spending to Medicare administration. For example, it notes that the Treasury Department collects taxes for Medicare and the Justice Department pursues fraud cases in the system. As the report explains to get to this number:

"Medicare unreported costs include parts of salaries for legislators, staff and others working on Medicare, building costs, marketing costs, collection of premiums and taxes, and accounting, including auditing and fraud issues, etc."

While these costs are not detailed carefully, Table 4 in the appendix shows that the study allocated 14.5 percent of the salaries of members of Congress and staff, as well executive branch salaries, to the administrative costs of Medicare. It also allocated 14.5 percent of non-correctional judicial costs.

Add a comment

Economists usually argue that it's best to tax the things you want discourage, like cigarettes, alcohol, and gasoline, not things you want to encourage, like work. That is why it is striking that that the Washington Post could not find one economist who thought that a plan in London to tax vacant housing units is a good idea. 

The only expert cited in the piece argued that the tax would have little effect on the housing market because the rich would not care if their property taxes were doubled or tripled, they would still leave units vacant. (The piece repeatedly refers to "experts" even though only one is cited by name.) This answer is striking for three reasons.

Whenever tax increases on the rich are proposed, the Washington Post and other newspapers are usually filled with assertions from economists about how it will cause to them work less, move, or take other steps to avoid the tax. Apparently for some reason a tax on vacant property is different from other taxes; the rich don't mind paying it.

The second reason the comment is striking is that it suggests a relatively costless way for the city of London to raise money. If doubling or tripling the property tax for vacant properties won't affect the behavior of the rich then maybe they could quadruple or quintuple the tax. Hey, if the rich don't mind paying a tax on vacant property, why not charge them ten or twenty times the regular property tax? They won't even notice.

The third reason is that the comment comes just before a statement telling us most vacant properties in London are owned by middle-class people, not rich people:

"Most of London’s 20,000 empty units aren’t owned by the super-rich, but by the middle and lower rungs — and studies show that most homes are empty because they are the subjects of inheritance tussles or uninhabitable or in need of repairs."

Taxing these people on vacant units might be just the incentive they need to resolve inheritance disputes or to sell a vacant property to someone who could afford to make the repairs for it to be habitable. Of course, this works from the assumption that we actually want less vacant property.

In arguing the case against the tax, the Post even found someone was happy to live next to a building with many vacant units:

"Plus, there are upsides to living next to absentee neighbors.

"'I would shudder to imagine if this was running at 80 percent occupancy,' said Suruchi Shukla, a 38-year-old consultant, with her head cranked skyward at the building."

So there you have it: a tax on vacant property wouldn't reduce the number of vacant units, and it would be a bad thing if it did. And remember folks, this is supposed to be a news story, not an opinion piece.

Add a comment

The NYT had an article on the most recent Republican plan to repeal the Affordable Care Act (ACA). It included a quote from Senator Bill Cassady, a co-sponsor of the bill:

"'Right now, 37 percent of the revenue from the Affordable Care Act goes to Americans in four states' — California, New York, Massachusetts and Maryland, Mr. Cassidy said. 'That is frankly not fair.'"

If just four states getting 37 percent of the money sounds unfair to you, it might be worth keeping in mind that these four states account for more than a quarter of the country's population and GDP. That still means that they are getting a disproportionate share of the money from the ACA, but it is not quite the same as if four small or average size states were getting 37 percent of the money.

It is also worth remembering that ACA spending accounts for roughly 1.5 percent of the total budget. Getting a disproportionate share of 1.5 percent of the total budget doesn't seem like all that big a deal, after all southern states have historically gotten a grossly disproportionate share of the military budget.

Then we have to consider the fact that many states don't get much money from the ACA by choice. Specifically, they have chosen not to expand Medicaid to provide more of their residents with health care insurance. They also have not done much to publicize the exchanges, which would have sent more money to their states through subsidies to moderate income people.

So this is the grave injustice that Senator Cassidy is looking to remedy and to do so he wants to have 30 million plus people lose their health care insurance, and to allow  insurers to again discriminate against people based on pre-existing conditions. That may be unfortunate, but at least California, New York, Massachusetts and Maryland won't be getting a disproportionate share of ACA spending.

Add a comment

We all know how hard it is for people with advanced degrees to compete in the global economy, but how much government help do we have to give them? Apparently, whatever we give them in strong patent and copyright protections (yes folks, this is protectionism, no matter how much you like it) is not enough.

In an editorial on relations with China, the NYT offers this encouraging (to the rich) pronouncement:

"On intellectual property, now that China is putting energy into developing its own technology instead of just stealing America’s, the two could work together on stronger protections."

Isn't that great? We can redistribute more money to people who benefit from patent and copyright monopolies and then say it's just unfortunate that technology leads to an upward redistribution of income. And most of our intellectual class are sufficiently s**t-for-brains to treat that as a serious argument.

And yes folks, there are alternatives to patent and copyright monopolies for financing research and creative work. But, as we all know, intellectuals have a hard time dealing with new ideas.

Add a comment

Of course, the paper did not actually warn of the cheap trick the Republicans are considering. It just mentioned it in passing as though it was a serious policy proposal:

"House Republicans are considering a key change to 401(k)s as part of their tax overhaul package: Taxing the money that workers place in their savings plans upfront instead of years later when they take it out in retirement."

The only possible rationale for taxing money when it is put into a 401(k) account rather than when it is pulled out is to change the timing of tax collections since there will be little net change in revenue over the long-term. By taxing the payments into the system, the government will collect more revenue during the 10-year horizon over which budget projections are made. However, this will be almost completely offset by lower tax collections in later years.

As policy, this makes zero sense. The point of 401(k)s is supposed to be encouraging people to save. There is much research showing that the prospect of an immediate tax saving gives strong incentive to save. This means that eliminating the immediate tax deduction will almost certainly mean less savings in 401(k)s.

There is the advantage that this change will appear to offset the lost revenue from Republican tax cuts for rich people. Unfortunately, most Post readers might not be aware of this rationale for the policy change.

Add a comment

Yeah, what else is new. After all, no one thinks that a conservative columnist writing for the country's most important newspaper should have any clue about the topics they cover.

Brooks uses his column today to tell us that workers are getting their share of the economic pie and the real problem is productivity growth.

"The problem of the middle-class squeeze, in short, may not be with how the fruits of productivity are distributed, but the fact that there isn’t much productivity growth at all. It’s not that a rising tide doesn’t lift all boats; it’s that the tide is not rising fast enough."

Sorry Mr. Brooks, but the numbers don't agree with you. Here's the story on average hourly earnings for production and non-supervisory workers since 2007. This is a good proxy for the median wage since it excludes high-end workers like doctors, Wall Street types, and CEOs.

wages production

Source: Bureau of Labor Statistics.

After rising by less than 2.0 percent between 2007 and 2013, real wages have started to grow at a respectable pace in the last four years. Still, they are just 8.0 percent above their 2007 level. By contrast, productivity has risen by 13.7 percent over this period. That's a difference of 5.7 percentage points over the last decade.

For a person earning $40,000 a year, the loss due to the gap between productivity growth and wage growth is equivalent to a tax increase of $2,280 a year. Would Brooks try to tell readers that a tax increase of this magnitude is small change? If we go further back to 1980, when inequality first started to take off, the gap would be closer to 40 percentage points. Anyhow, it's cute that Brooks wants to tell us not to think about all the money going to rich people and just concentrate on productivity growth, but this stuff is too silly even for children's games.

There, however, is one point here worth noting. As unemployment has fallen, workers have been seeing their share of gains from productivity growth. This is a huge deal, which points to the importance of Federal Reserve Board policy.

There were many economists who wanted the Fed to take a tighter stance on monetary policy and keep the unemployment rate from falling as low as it has. The decision by the Fed to be relatively slow in raising interest rates has not only allowed millions of additional workers to get jobs, it has also meant higher pay for those who were already working. This makes an enormous difference to the country's workers, but don't expect to see anything about it in a David Brooks column.

Add a comment

Ruchir Sharma, the chief global strategist at Morgan Stanley Investment Management, used his NYT column to argue that central banks have to include fighting asset bubbles on their agenda, in addition to promoting high employment and low inflation. As someone who has argued this for two decades, I am sympathetic to the point; however, Sharma gets a couple of big things wrong.

First, the big issue with bubbles is whether they are moving the economy. This is something that is easy to determine for folks familiar with introductory economics. The issue here is whether some component of demand is out of line with its long-term trend.

That was easy to see in the late 1990s as the wealth created by the stock bubble led to a consumption boom, pushing the saving rate to a then-record low. The investment share of GDP also became unusually high, with investment concentrated in the tech sector where stock prices were most out of line with corporate profits. 

The same was true of the housing bubble in the last decade. Residential construction hit a record 6.5 percent share of GDP, a level that clearly did not make sense given the underlying demographics of the country. The wealth effect from the bubble created housing wealth led to an even larger consumption boom than the 1990s stock bubble, as savings rates fell even lower than they had in the late 1990s. It is difficult to understand how the Fed could have missed the impact of the bubble or think that these sources of demand could be easily replaced when the bubble burst.

Add a comment

Morning Edition had a segment on efforts to counter the Trump administration's proposal to cut foreign aid, including money going to the Global Fund to Fight AIDS, Tuberculosis, and Malaria. The segment featured a comment from Bill Gates, whose foundation also supports this fund.

Gates noted the lack of support for foreign aid and attributed it to the media's tendency to highlight failures, where money is poorly spent or stolen, as opposed to the success stories. While there is undoubtedly some truth to this argument, the more likely problem is the fratboy reporting on budget issues that gives the audience zero context.

In the case of the Global Fund, when people hear about it they are likely to hear that the U.S. will spend up to $4.3 billion this year. This might lead people to think that we are devoting a substantial share of the budget to this fund, at the expense of other programs and/or raising people's taxes.

In fact, this sum is roughly to 0.1 percent of the budget, it comes to roughly $13 per person for everyone in the country. For another comparison, it is roughly 36 times the excess cost (the amount paid beyond what it cost to protect a normal president) of Donald Trump's secret service protection. In other words, it would not be the reason that people are paying taxes they consider high.

Polls consistently show the public grossly overestimates the share of the budget going to foreign aid, with median estimates in the range of 25 to 30 percent of the budget, when the total is less than 1.0 percent, even when being very generous about what counts as aid. This is likely due at least in part to the fratboy reporting on aid, which gives people no information whatsoever.

It is amazing that National Public Radio, the New York Times, and other leading news outlets continue the bizarre practice of reporting large numbers without context, even when everyone knows it is not providing information to the vast majority of their audience. It is a very simple matter to express a budget number as a share of the total budget or to use some other comparison that would give it meaning to listeners and readers. However, news outlets refuse to do this.

They instead engage in mindless fratboy reporting where they carry through the ritual of giving a large number that means absolutely nothing to the people who see it, and then pretend they have done their job. This is a major reason people are hostile to programs like foreign aid, TANF, food stamps and other programs that benefit the poor here and elsewhere.

Yes, I know many people are racist and hate these programs for that reason. But many people who are not especially racist (i.e. they vote for people like Hillary Clinton) also believe that 30 percent of our budget goes to foreign aid or TANF. If these programs actually did cost that much much money, there would be good reasons for not supporting them, since they don't have that much to show.

Add a comment

The Census Bureau reported that the percentage of people without health insurance fell by 0.3 percentage points in 2016 to 8.8 percent. This puts the cumulative gain in coverage since 2013, when the main Affordable Care Act provisions took effect, at 4.5 percentage points.

By state, the largest drop in the percent of the population that is uninsured was in California, which had a decline of 9.8 percentage points to 7.3 percent. Next was New Mexico with a decline of 9.5 percentage points, giving it an uninsurance rate of 9.2 percent, and Nevada with a drop of 9.3 percentage points to 11.4 percent.

The smallest decline over this period was in Massachusetts, where the percent uninsured fell by just 1.2 percentage points, but this is due to the fact that it had an uninsured rate of just 3.7 percent in 2013. Wyoming had the second smallest decline, with the rate falling by 1.3 percentage points to 11.5 percent.

Texas, Alaska, and Oklahoma had the highest uninsured rate in 2016, at 16.6 percent, 14.0 percent, and 13.8 percent, respectively. The lowest rates were in Massachusetts, 2.5 percent, Hawaii, 3.5 percent, and Vermont, 3.7 percent.

By employment, the biggest increase in coverage was among those working part-time. The percentage uninsured among this group fell by 9.0 percentage points to 14.8 percent. For those who did not work at all, the percent of uninsured fell by 7.4 percentage points to 15.0 percent. For full-time, full-year, workers the drop was 4.1 percentage points to 9.8 percent.

On the whole, these newest numbers indicate that Obamacare has succeeded somewhat more than expected in extending coverage. The biggest beneficiaries have been people who choose to work part-time (80 percent of part-time employment is voluntary), who no longer need to get coverage through an employer as a result of the exchanges and the expansion of Medicaid.

 

Note: Percent of uninsured for Wyoming has been corrected; thanks, Charles Angevine.

Add a comment

According to an article in Sunday's paper, employers are now able to find the workers they need by going to cities with large amounts of unemployment or underemployment.

Add a comment

Readers may have missed this fact, but this is what the NYT said in an article on the prospect for tax reform when it told readers:

"Democrats have also been deeply skeptical of the Trump administration’s plans to repeal the estate tax, which it has said has been harmful to family farmers."

If the Trump administration has been saying the estate tax has been harmful to family farmers it is lying since virtually no family farmers will owe a penny as a result of the estate tax. This would be known to NYT since the paper ran a piece 16 years ago which noted that the American Farm Bureau Federation, a strong opponent of the estate tax, could not identify a single person who had lost a family farm as a result of the estate tax.

Add a comment