The size and position of an element are often impacted by its containing block. Most often, the containing block is the content area of an element's nearest block-level ancestor, but this is not always the case.
When a user agent (such as your browser) lays out a document, it generates a box for every element. Each box is divided into four areas:
- Content area
- Padding area
- Border area
- Margin area

Many developers believe that the containing block of an element is always the content area of its parent. But this is false! Read on...
Why is it important to identify the containing block?
The size and position of an element are often impacted by its containing block. Percentage values that are applied to the width, height, padding, margin, and offset properties of an absolutely positioned element (i.e., which has its position set to absolute or fixed) are computed from the element's containing block.
How can we identify the containing block?
The process for identifying the containing block depends entirely on the value of the element's position property:
- If the
positionproperty isstaticorrelative, the containing block is formed by the edge of the content box of the nearest ancestor element that is a block container (such as an inline-block, block, or list-item element) or which establishes a formatting context (such as a table container, flex container, grid container, or the block container itself). - If the
positionproperty isabsolute, the containing block is formed by the edge of the padding box of the nearest ancestor element that has apositionvalue other thanstatic(fixed,absolute,relative, orsticky). - If the
positionproperty isfixed, the containing block is established by the viewport (in the case of continuous media) or the page area (in the case of paged media). - If the
positionproperty isabsoluteorfixed, the containing block may also be formed by the edge of the padding box of the nearest ancestor element that has the following:- A
transformorperspectivevalue other thannone - A
will-changevalue oftransformorperspective - A
filtervalue other thannoneor awill-changevalue offilter(only works on Firefox).
- A
Note: The containing block in which the root element (<html>) resides is a rectangle called the initial containing block. It has the dimensions of the viewport (for continuous media) or the page area (for paged media).
Calculating percentage values from the containing block
As noted above, when certain properties are given a percentage value, the computed value depends on the element's containing block. The properties that work this way are box model properties and offset properties:
- The
height,top, andbottomproperties compute percentage values from the height of the containing block. If the height of the containing block depends on its contents, these values become0when the containing block has apositionofrelativeorstatic. - The
width,left,right,padding, andmarginproperties compute percentage values from the width of the containing block.
Some examples
The HTML code for all our examples is:
<body>
<section>
<p>This is a paragraph!</p>
</section>
</body>
Example 1
In this example, the paragraph is statically positioned, so its containing block is <section> because it's the nearest ancestor that is a block container.
<body>
<section>
<p>This is a paragraph!</p>
</section>
</body>
body {
background: beige;
}
section {
display: block;
width: 400px;
height: 160px;
background: lightgray;
}
p {
width: 50%; /* == 400px * .5 = 200px */
height: 25%; /* == 160px * .25 = 40px */
margin: 5%; /* == 400px * .05 = 20px */
padding: 5%; /* == 400px * .05 = 20px */
background: cyan;
}
Example 2
In this example, the paragraph's containing block is the <body> element, because <section> is not a block container and doesn’t establish a formatting context.
<body>
<section>
<p>This is a paragraph!</p>
</section>
</body>
body {
background: beige;
}
section {
display: inline;
background: lightgray;
}
p {
width: 50%; /* == half the body's width */
height: 200px; /* Note: a percentage would be 0 */
background: cyan;
}
Example 3
In this example, the paragraph's containing block is <section> because the latter's position is absolute. The paragraph's percentage values are affected by the padding of its containing block, though if the containing block's box-sizing value were border-box this would not be the case.
<body>
<section>
<p>This is a paragraph!</p>
</section>
</body>
body {
background: beige;
}
section {
position: absolute;
left: 30px;
top: 30px;
width: 400px;
height: 160px;
padding: 30px 20px;
background: lightgray;
}
p {
position: absolute;
width: 50%; /* == (400px + 20px + 20px) * .5 = 220px */
height: 25%; /* == (160px + 30px + 30px) * .25 = 55px */
margin: 5%; /* == (400px + 20px + 20px) * .05 = 22px */
padding: 5%; /* == (400px + 20px + 20px) * .05 = 22px */
background: cyan;
}
Example 4
In this example, the paragraph's position is fixed, so its containing block is the initial containing block (on screens, the viewport). Thus, the paragraph's dimensions change based on the size of the browser window.
<body>
<section>
<p>This is a paragraph!</p>
</section>
</body>
body {
background: beige;
}
section {
width: 400px;
height: 480px;
margin: 30px;
padding: 15px;
background: lightgray;
}
p {
position: fixed;
width: 50%; /* == (50vw - (width of vertical scrollbar)) */
height: 50%; /* == (50vh - (height of horizontal scrollbar)) */
margin: 5%; /* == (5vw - (width of vertical scrollbar)) */
padding: 5%; /* == (5vw - (width of vertical scrollbar)) */
background: cyan;
}
Example 5
In this example, the paragraph's position is absolute, so its containing block is <section>, which is the nearest ancestor with a transform property that isn't none.
<body>
<section>
<p>This is a paragraph!</p>
</section>
</body>
body {
background: beige;
}
section {
transform: rotate(0deg);
width: 400px;
height: 160px;
background: lightgray;
}
p {
position: absolute;
left: 80px;
top: 30px;
width: 50%; /* == 200px */
height: 25%; /* == 40px */
margin: 5%; /* == 20px */
padding: 5%; /* == 20px */
background: cyan;
}